This is an internal, incomplete preview of a proposed change to the ACL Anthology.
For efficiency reasons, we generate only three BibTeX files per volume, and the preview may be incomplete in other ways, or contain mistakes.
Do not treat this content as an official publication.
Large language models (LLMs) can perform complex reasoning by generating intermediate reasoning steps using chain-of-thought prompting under zero-shot or few-shot settings. However, zero-shot prompting always encounters low performance, and the superior performance of few-shot prompting hinges on the manual-crafting of task-specific demonstrations one by one. In this paper, we present **RoSE** (**R**easoning with **O**rchestrated **S**treaming **E**xperiences), a general framework for solving reasoning tasks that can self-improve as it answers various reasoning questions. To enable RoSE, we describe an architecture that extends an LLM to store all answered reasoning questions and their reasoning steps in a streaming experience pool and orchestrate helpful questions from the pool to assist itself in answering new questions. To set up a question-aware orchestration mechanism, RoSE first calculates the similarity of each question in the pool with the question to be answered. Since the solution to each question in the experience pool is not always correct, RoSE will sort the questions according to their similarity with the question to be answered, and then uniformly divide them into multiple buckets. It finally extracts one question from each bucket to make the extracted questions more diverse. To make the extracted questions help RoSE answer new questions as much as possible, we introduce two other attributes of uncertainty and complexity for each question. RoSE will preferentially select the questions with low uncertainty and high complexity from each bucket. We evaluate the versatility of RoSE in various complex reasoning tasks and LLMs, such as arithmetic and commonsense reasoning, and find that it can achieve excellent performance without any labeled data and pre-set unlabeled data.
Large language models optimized with techniques like RLHF have achieved good alignment in being helpful and harmless. However, post-alignment, these language models often exhibit overconfidence, where the expressed confidence does not accurately calibrate with their correctness rate. In this paper, we decompose the language model confidence into the Uncertainty about the question and the Fidelity to the answer generated by language models. Then, we propose a plug-and-play method, UF Calibration, to estimate the confidence of language models. Our method has shown good calibration performance by conducting experiments with 6 RLHF-LMs on four MCQA datasets. Moreover, we propose two novel metrics, IPR and CE, to evaluate the calibration of the model, and we have conducted a detailed discussion on Truly Well-Calibrated Confidence for large language models. Our method could serve as a strong baseline, and we hope that this work will provide some insights into the model confidence calibration.
As large language models (LLMs) rapidly evolve, they are increasingly being customized through fine-tuning to suit the specific needs of various applications. A critical aspect of this advancement is the alignment process, which ensures that these models perform tasks in ways that align with human values and expectations. Current alignment methods, such as direct preference optimization (DPO) and reinforcement learning from human feedback (RLHF), focus primarily on alignment during training phase. However, these methods often involve complex and resource-intensive training processes, posing significant challenge for their implementation. Therefore, we propose InferAligner, a simple yet effective method for harmlessness alignment during inference phase. InferAligner decouples harmlessness from helpfulness. During the training phase, it focuses solely on enhancing the target model’s capabilities on downstream tasks. In the inference phase, it utilizes safety steering vectors extracted from the aligned model to guide the target model towards harmlessness alignment. Experimental results show that our method can be very effectively applied to domain-specific models in finance, medicine, and mathematics, as well as to multimodal large language models (MLLMs) such as LLaVA. It significantly diminishes the attack success rate (ASR) of both harmful instructions and jailbreak instructions, while maintaining almost unchanged performance in downstream tasks.
Sparse Mixture of Experts (MoE) models are popular for training large language models due to their computational efficiency. However, the commonly used top-k routing mechanism suffers from redundancy computation and memory costs due to the unbalanced routing. Some experts are overflow, where the exceeding tokens are dropped. While some experts are empty, which are padded with zeros, negatively impacting model performance. To address the dropped tokens and padding, we propose the Rectify-Router, comprising the Intra-GPU Rectification and the Fill-in Rectification. The Intra-GPU Rectification handles dropped tokens, efficiently routing them to experts within the GPU where they are located to avoid inter-GPU communication. The Fill-in Rectification addresses padding by replacing padding tokens with the tokens that have high routing scores. Our experimental results demonstrate that the Intra-GPU Rectification and the Fill-in Rectification effectively handle dropped tokens and padding, respectively. Furthermore, the combination of them achieves superior performance, surpassing the accuracy of the vanilla top-1 router by 4.7%.
The evolution of Large Language Models (LLMs) has led to significant advancements, with models like Claude and Gemini capable of processing contexts up to 1 million tokens. However, efficiently handling long sequences remains challenging, particularly during the prefilling stage when input lengths exceed GPU memory capacity. Traditional methods often segment sequence into chunks and compress them iteratively with fixed-size memory. However, our empirical analysis shows that the fixed-size memory results in wasted computational and GPU memory resources. Therefore, we introduces Incremental Memory (IM), a method that starts with a small memory size and gradually increases it, optimizing computational efficiency. Additionally, we propose Decremental Chunk based on Incremental Memory (IMDC), which reduces chunk size while increasing memory size, ensuring stable and lower GPU memory usage. Our experiments demonstrate that IMDC is consistently faster (1.45x) and reduces GPU memory consumption by 23.3% compared to fixed-size memory, achieving comparable performance on the LongBench Benchmark.
In the realm of Large Language Models (LLMs), users commonly employ diverse decoding strategies and adjust hyperparameters to control the generated text. However, a critical question emerges: Are LLMs conscious of the existence of these decoding strategies and capable of regulating themselves? The current decoding generation process often relies on empirical and heuristic manual adjustments to hyperparameters based on types of tasks and demands. However, this process is typically cumbersome, and the decoding hyperparameters may not always be optimal for each sample. To address the aforementioned challenges, we propose a novel text generation paradigm termed Hyperparameter Aware Generation (HAG). By leveraging hyperparameter-aware instruction tuning, the LLM autonomously determines the optimal decoding strategy and configs based on the input samples, enabling self-regulation. Our approach eliminates the need for extensive manual tuning, offering a more autonomous, self-regulate model behavior. Experimental results spanning six datasets across reasoning, creativity, translation, and mathematics tasks demonstrate that hyperparameter-aware instruction tuning empowers the LLMs to self-regulate the decoding strategy and hyperparameter. HAG extends the current paradigm in the text generation process, highlighting the feasibility of endowing the LLMs with self-regulate decoding strategies.
Although large language models (LLMs) have demonstrated remarkable performance, the lack of transparency in their inference logic raises concerns about their trustworthiness. To gain a better understanding of LLMs, we conduct a detailed analysis of the operations of attention heads and aim to better understand the in-context learning of LLMs. Specifically, we investigate whether attention heads encode two types of relationships between tokens present in natural languages: the syntactic dependency parsed from sentences and the relation within knowledge graphs. We find that certain attention heads exhibit a pattern where, when attending to subject tokens, they recall object tokens and increase the output logits of those object tokens. More crucially, the formulation of such semantic induction heads has a close correlation with the emergence of the in-context learning ability of language models. The study of semantic attention heads advances our understanding of the intricate operations of attention heads in transformers, and further provides new insights into the in-context learning of LLMs.
The Large Vision-Language Models (LVLMs) have demonstrated great abilities in image perception and language understanding. However, existing datasets either focus solely on primary perception abilities and commonsense knowledge, or have a low level of text comprehension difficulty, which are insufficient to reflect the comprehensive capabilities of LVLMs, particularly in terms of Chinese language proficiency. We propose GAOKAO-MM, a multimodal benchmark based on the Chinese College Entrance Examination (GAOKAO), comprising of 8 subjects and 12 types of images, such as diagrams, function graphs, maps and photos. GAOKAO-MM derives from native Chinese context and sets human-level requirements for the model’s abilities, including perception, understanding, knowledge and reasoning. We evaluate 10 LVLMs and find that the accuracies of all of them are lower than 50%, with GPT-4-Vision (48.1%), Qwen-VL-Plus (41.2%) and Gemini-Pro-Vision (35.1%) ranking in the top three positions. The results of our multi-dimension analysis indicate that LVLMs have moderate distance towards Artificial General Intelligence (AGI) and provide insights facilitating the development of multilingual LVLMs. The dataset and evaluation code are available through: https://github.com/OpenMOSS/GAOKAO-MM
Large language models have achieved remarkable success, but their extensive parameter size necessitates substantial memory for training, thereby setting a high threshold. While the recently proposed low-memory optimization (LOMO) reduces memory footprint, its optimization technique, akin to stochastic gradient descent, is sensitive to hyper-parameters and exhibits suboptimal convergence, failing to match the performance of the prevailing optimizer for large language models, AdamW. Through analysis of the Adam optimizer, we found that, compared to momentum, the adaptive learning rate is more critical for bridging the gap. Building on this insight, we introduce the low-memory optimization with adaptive learning rate (AdaLomo), which offers an adaptive learning rate for each parameter and exhibits superior convergence performance compared to LOMO theoretically. To maintain memory efficiency, we employ non-negative matrix factorization for the second-order moment estimation. Additionally, we suggest the use of a grouped update normalization to stabilize convergence. Our experiments with instruction-tuning and further pre-training demonstrate that AdaLomo achieves results on par with AdamW, while significantly reducing memory requirements, thereby lowering the hardware barrier to training large language models. The code is accessible at https://github.com/OpenLMLab/LOMO.
The programming skill is one crucial ability for Large Language Models (LLMs), necessitating a deep understanding of programming languages (PLs) and their correlation with natural languages (NLs). We examine the impact of pre-training data on code-focused LLMs’ performance by assessing the comment density as a measure of PL-NL alignment. Given the scarcity of code-comment aligned data in pre-training corpora, we introduce a novel data augmentation method that generates comments for existing code, coupled with a data filtering strategy that filters out code data poorly correlated with natural language. We conducted experiments on three code-focused LLMs and observed consistent improvements in performance on two widely-used programming skill benchmarks. Notably, the model trained on the augmented data outperformed both the model used for generating comments and the model further trained on the data without augmentation.
Data plays a fundamental role in the training of Large Language Models (LLMs). While attention has been paid to the collection and composition of datasets, determining the data sampling strategy in training remains an open question. Most LLMs are trained with a simple strategy, random sampling. However, this sampling strategy ignores the unbalanced nature of training data distribution, which can be sub-optimal. In this paper, we propose ClusterClip Sampling to balance the text distribution of training data for better model training. Specifically, ClusterClip Sampling utilizes data clustering to reflect the data distribution of the training set and balances the common samples and rare samples during training based on the cluster results. A repetition clip operation is introduced to mitigate the overfitting issue led by samples from certain clusters. Extensive experiments validate the effectiveness of ClusterClip Sampling, which outperforms random sampling and other cluster-based sampling variants under various training datasets and large language models.
The quality of training data is crucial for enhancing the long-text capabilities of foundation models. Despite existing efforts to refine data quality through heuristic rules and evaluations based on data diversity and difficulty, there’s a lack of systematic approaches specifically tailored for assessing long texts. Addressing this gap, our work systematically measures the quality of long texts by evaluating three fundamental linguistic dimensions: coherence, cohesion, and complexity. Drawing inspiration from the aforementioned three dimensions, we introduce a suite of metrics designed to evaluate the quality of long texts, encompassing both statistical and pre-trained language model-based ones. Leveraging these metrics, we present LongWanjuan, a bilingual dataset specifically tailored to enhance the training of language models for long-text tasks with over 160B tokens. In LongWanjuan, we categorize long texts into holistic, aggregated, and chaotic types, enabling a detailed analysis of long-text quality. Furthermore, we devise a data mixture recipe that strategically balances different types of long texts within LongWanjuan, leading to significant improvements in model performance on long-text tasks.
The training process of large language models (LLMs) often involves varying degrees of test data contamination. Although current LLMs are achieving increasingly better performance on various benchmarks, their performance in practical applications does not always match their benchmark results. Leakage of benchmarks can prevent the accurate assessment of LLMs’ true performance. However, constructing new benchmarks is costly, labor-intensive and still carries the risk of leakage. Therefore, in this paper, we ask the question Can we reuse these leaked benchmarks for LLM evaluation? We propose Inference-Time Decontamination (ITD) to address this issue by detecting and rewriting leaked samples without altering their difficulties. ITD can mitigate performance inflation caused by memorizing leaked benchmarks. Our proof-of-concept experiments demonstrate that ITD reduces inflated accuracy by 22.9% on GSM8K and 19.0% on MMLU. On MMLU, using Inference-time Decontamination can lead to a decrease in the results of Phi3 and Mistral by 6.7% and 3.6% respectively. We hope that ITD can provide more truthful evaluation results for large language models.
Fact knowledge memorization is crucial for Large Language Models (LLM) to generate factual and reliable responses. However, the behaviors of LLM fact memorization remain under-explored. In this paper, we analyze the scaling laws for LLM’s fact knowledge and LLMs’ behaviors of memorizing different types of facts. We find that LLMs’ fact knowledge capacity has a linear and negative exponential law relationship with model size and training epochs, respectively. Estimated by the built scaling law, memorizing the whole Wikidata’s facts requires training an LLM with 1000B non-embed parameters for 100 epochs, suggesting that using LLMs to memorize all public facts is almost implausible for a general pre-training setting. Meanwhile, we find that LLMs can generalize on unseen fact knowledge and its scaling law is similar to general pre-training. Additionally, we analyze the compatibility and preference of LLMs’ fact memorization. For compatibility, we find LLMs struggle with memorizing redundant facts in a unified way. Only when correlated facts have the same direction and structure, the LLM can compatibly memorize them. This shows the inefficiency of LLM memorization for redundant facts. For preference, the LLM pays more attention to memorizing more frequent and difficult facts, and the subsequent facts can overwrite prior facts’ memorization, which significantly hinders low-frequency facts memorization. Our findings reveal the capacity and characteristics of LLMs’ fact knowledge learning, which provide directions for LLMs’ fact knowledge augmentation.
While current tasks of converting natural language to SQL (NL2SQL) using Foundation Models have shown impressive achievements, adapting these approaches for converting natural language to Graph Query Language (NL2GQL) encounters hurdles due to the distinct nature of GQL compared to SQL, alongside the diverse forms of GQL. Moving away from traditional rule-based and slot-filling methodologies, we introduce a novel approach, R3-NL2GQL, integrating both small and large Foundation Models for ranking, rewriting, and refining tasks. This method leverages the interpretative strengths of smaller models for initial ranking and rewriting stages, while capitalizing on the superior generalization and query generation prowess of larger models for the final transformation of natural language queries into GQL formats. Addressing the scarcity of datasets in this emerging field, we have developed a bilingual dataset, sourced from graph database manuals and selected open-source Knowledge Graphs (KGs). Our evaluation of this methodology on this dataset demonstrates its promising efficacy and robustness.
In Retrieval-Augmented Generation (RAG), retrieval is not always helpful and applying it to every instruction is sub-optimal. Therefore, determining whether to retrieve is crucial for RAG, which is usually referred to as Active Retrieval. However, existing active retrieval methods face two challenges: 1. They usually rely on a single criterion, which struggles with handling various types of instructions. 2. They depend on specialized and highly differentiated procedures, and thus combining them makes the RAG system more complicated and leads to higher response latency. To address these challenges, we propose Unified Active Retrieval (UAR). UAR contains four orthogonal criteria and casts them into plug-and-play classification tasks, which achieves multifaceted retrieval timing judgements with negligible extra inference cost. We further introduce the Unified Active Retrieval Criteria (UAR-Criteria), designed to process diverse active retrieval scenarios through a standardized procedure. Experiments on four representative types of user instructions show that UAR significantly outperforms existing work on the retrieval timing judgement and the performance of downstream tasks, which shows the effectiveness of UAR and its helpfulness to downstream tasks.
The widespread adoption of large language models (LLMs) across various regions underscores the urgent need to evaluate their alignment with human values. Current benchmarks, however, fall short of effectively uncovering safety vulnerabilities in LLMs. Despite numerous models achieving high scores and ‘topping the chart’ in these evaluations, there is still a significant gap in LLMs’ deeper alignment with human values and achieving genuine harmlessness. To this end, this paper proposes a value alignment benchmark named Flames, which encompasses both common harmlessness principles and a unique morality dimension that integrates specific Chinese values such as harmony. Accordingly, we carefully design adversarial prompts that incorporate complex scenarios and jailbreaking methods, mostly with implicit malice. By prompting 17 mainstream LLMs, we obtain model responses and rigorously annotate them for detailed evaluation. Our findings indicate that all the evaluated LLMs demonstrate relatively poor performance on Flames, particularly in the safety and fairness dimensions. We also develop a lightweight specified scorer capable of scoring LLMs across multiple dimensions to efficiently evaluate new models on the benchmark. The complexity of Flames has far exceeded existing benchmarks, setting a new challenge for contemporary LLMs and highlighting the need for further alignment of LLMs. Our benchmark is publicly available at https://github.com/AIFlames/Flames.
Verifiable generation aims to let the large language model (LLM) generate text with supporting documents, which enables the user to flexibly verify the answer and makes the LLM’s output more reliable. Retrieval plays a crucial role in verifiable generation. Specifically, the retrieved documents not only supplement knowledge to help the LLM generate correct answers, but also serve as supporting evidence for the user to verify the LLM’s output. However, the widely used retrievers become the bottleneck of the entire pipeline and limit the overall performance. Their capabilities are usually inferior to LLMs since they often have much fewer parameters than the large language model and have not been demonstrated to scale well to the size of LLMs. If the retriever does not correctly find the supporting documents, the LLM can not generate the correct and verifiable answer, which overshadows the LLM’s remarkable abilities. To address these limitations, we propose **LLatrieval** (**L**arge **La**nguage Model Verified Re**trieval**),where the LLM updates the retrieval result until it verifies that the retrieved documents can sufficiently support answering the question. Thus, the LLM can iteratively provide feedback to retrieval and facilitate the retrieval result to fully support verifiable generation. Experiments on ALCE show that LLatrieval significantly outperforms extensive baselines and achieves state-of-the-art results.
Machine reasoning, which involves solving complex problems through step-by-step deduction and analysis, is a crucial indicator of the capabilities of Large Language Models (LLMs). However, as the complexity of tasks escalates, LLMs often encounter increasing errors in their multi-step reasoning process. This study delves into the underlying factors contributing to these reasoning errors and seeks to leverage uncertainty to refine them. Specifically, we introduce Uncertainty-aware Adaptive Guidance (UAG), a novel approach for guiding LLM reasoning onto an accurate and reliable trajectory. UAG first identifies and evaluates uncertainty signals within each step of the reasoning chain. Upon detecting a significant increase in uncertainty, UAG intervenes by retracting to a previously reliable state and then introduces certified reasoning clues for refinement. By dynamically adjusting the reasoning process, UAG offers a plug-and-play solution for improving LLMs’ performance in complex reasoning. Extensive experiments across various reasoning tasks demonstrate that UAG not only enhances the reasoning abilities of LLMs but also consistently outperforms several strong baselines with minimal computational overhead. Further analysis reveals that UAG is notably effective in identifying and diminishing reasoning errors.
Reconstructing natural language from non-invasive electroencephalography (EEG) holds great promise as a language decoding technology for brain-computer interfaces (BCIs). However, EEG-based language decoding is still in its nascent stages, facing several technical issues such as: 1) Absence of a hybrid strategy that can effectively integrate cross-modality (between EEG and text) self-learning with intra-modality self-reconstruction of EEG features or textual sequences; 2) Under-utilization of large language models (LLMs) to enhance EEG-based language decoding. To address above issues, we propose the Contrastive EEG-Text Masked Autoencoder (CET-MAE), a novel model that orchestrates compound self-supervised learning across and within EEG and text through a dedicated multi-stream encoder. Furthermore, we develop a framework called E2T-PTR (EEG-to-Text decoding using Pretrained Transferable Representations), which leverages pre-trained modules alongside the EEG stream from CET-MAE and further enables an LLM (specifically BART) to decode text from EEG sequences. Comprehensive experiments conducted on the popular text-evoked EEG database, ZuCo, demonstrate the superiority of E2T-PTR, which outperforms the baseline framework in ROUGE-1 F1 and BLEU-4 scores by 8.34% and 32.21%, respectively. Our proposed pre-trained EEG-Text model shows the potential to improve downstream tasks involving EEG and text. This opens up promising avenues for its application in inner speech BCI paradigms, meriting further investigation.
Large Language Models (LLMs) have revolutionized Natural Language Processing (NLP) but demand massive GPU resources for training. Lowering the threshold for LLMs training would encourage greater participation from researchers, benefiting both academia and society. While existing approaches have focused on parameter-efficient fine-tuning, which tunes or adds a small number of parameters, few have addressed the challenge of tuning the full parameters of LLMs with limited resources. In this work, we propose a new optimizer, LOw-Memory Optimization (LOMO), which fuses the gradient computation and the parameter update in one step to reduce memory usage. By integrating LOMO with existing memory saving techniques, we reduce memory usage to 10.8% compared to the standard approach (DeepSpeed solution). Consequently, our approach enables the full parameter fine-tuning of a 65B model on a single machine with 8 × RTX 3090, each with 24GB memory. Code and data are available at https://github.com/OpenLMLab/LOMO.
Large language models (LLMs) garner significant attention for their unprecedented performance, leading to an increasing number of researches evaluating LLMs. However, these evaluation benchmarks are limited to assessing the instruction-following capabilities, overlooking the fundamental abilities that emerge during the pre-training stage. Previous subjective evaluation methods mainly reply on scoring by API models. However, in the absence of references, large models have shown limited ability to discern subtle differences. To bridge the gap, we propose F-Eval, a bilingual evaluation benchmark to evaluate the fundamental abilities, including expression, commonsense and logic. The tasks in F-Eval include multi-choice objective tasks, open-ended objective tasks, reference-based subjective tasks and reference-free subjective tasks. For reference-free subjective tasks, we devise new evaluation methods, serving as alternatives to scoring by API models. We conduct evaluations on 13 advanced LLMs. Results show that our evaluation methods show higher correlation coefficients and larger distinction than other evaluators. Additionally, we discuss the influence of different model sizes, dimensions, and normalization methods. We anticipate that F-Eval will facilitate the study of LLMs’ fundamental abilities.
We introduce AnyGPT, an any-to-any multimodal language model that utilizes discrete representations for the unified processing of various modalities, including speech, text, images, and music. AnyGPT can be trained stably without any alterations to the current large language model (LLM) architecture or training paradigms. Instead, it relies exclusively on data-level preprocessing, facilitating the seamless integration of new modalities into LLMs, akin to the incorporation of new languages.We build a multimodal text-centric dataset for multimodal alignment pre-training. Utilizing generative models, we synthesize the first large-scale any-to-any multimodal instruction dataset. It consists of 108k samples of multi-turn conversations that intricately interweave various modalities, thus equipping the model to handle arbitrary combinations of multimodal inputs and outputs.Experimental results demonstrate that AnyGPT is capable of facilitating any-to-any multimodal conversation while achieving performance comparable to specialized models across all modalities, proving that discrete representations can effectively and conveniently unify multiple modalities within a language model. Demos are shown in https://junzhan2000.github.io/AnyGPT.github.io/.
Recently, there has been growing interest in long-context scaling of large language models (LLMs). To facilitate research in this field, we propose L-Eval to institute a more standardized evaluation for Long-Context Language Models (LCLMs) addressing two key aspects: dataset construction and evaluation metrics. On the one hand, we build a new evaluation suite containing 20 sub-tasks, 508 long documents, and more than 2,000 human-labeled query-response pairs including diverse task types, domains, and input length (3k~200k tokens). On the other hand, we investigate the effectiveness of evaluation metrics for LCLMs and we show that Length-instruction-enhanced (LIE) evaluation and LLM judges can better correlate with human judgments. We conducted a comprehensive study of 4 popular commercial LLMs and 12 open-source counterparts using the L-Eval benchmark. Our empirical findings offer useful insights into the study of LCLMs and lay the groundwork for the development of a more principled evaluation of these models.
Recent advancements in Chain-of-Thought prompting have facilitated significant breakthroughs for Large Language Models (LLMs) in complex reasoning tasks. Current research enhances the reasoning performance of LLMs by sampling multiple reasoning chains and ensembling based on the answer frequency. However, this approach fails in scenarios where the correct answers are in the minority. We identify this as a primary factor constraining the reasoning capabilities of LLMs, a limitation that cannot be resolved solely based on the predicted answers. To address this shortcoming, we introduce a hierarchical reasoning aggregation framework AoR (Aggregation of Reasoning), which selects answers based on the evaluation of reasoning chains. Additionally, AoR incorporates dynamic sampling, adjusting the number of reasoning chains in accordance with the complexity of the task. Experimental results on a series of complex reasoning tasks show that AoR outperforms prominent ensemble methods. Further analysis reveals that AoR not only adapts various LLMs but also achieves a superior performance ceiling when compared to current methods.
Large language models (LLMs) are highly effective in various natural language processing (NLP) tasks. However, they are susceptible to producing unreliable conjectures in ambiguous contexts called hallucination. This paper presents a new method for evaluating LLM hallucination in Question Answering (QA) based on the unanswerable math word problem (MWP). To support this approach, we innovatively develop a dataset called Unanswerable Math Word Problem (UMWP) which comprises 5200 questions across five categories. We developed an evaluation methodology combining text similarity and mathematical expression detection to determine whether LLM considers the question unanswerable. The results of extensive experiments conducted on 31 LLMs, including GPT-3, InstructGPT, LLaMA, and Claude, demonstrate that in-context learning and reinforcement learning with human feedback (RLHF) training significantly enhance the model’s ability to avoid hallucination. We show that utilizing MWP is a reliable and effective approach to assess hallucination. Our code and data are available at https://github.com/Yuki-Asuuna/UMWP.
Most existing methods of Out-of-Domain (OOD) intent classification rely on extensive auxiliary OOD corpora or specific training paradigms. However, they are underdeveloped in the underlying principle that the models should have differentiated confidence in In- and Out-of-domain intent. In this work, we shed light on the fundamental cause of model overconfidence on OOD and demonstrate that calibrated subnetworks can be uncovered by pruning the overparameterized model. Calibrated confidence provided by the subnetwork can better distinguish In- and Out-of-domain, which can be a benefit for almost all post hoc methods. In addition to bringing fundamental insights, we also extend the Lottery Ticket Hypothesis to open-world scenarios. We conduct extensive experiments on four real-world datasets to demonstrate our approach can establish consistent improvements compared with a suite of competitive baselines.
Adversarial purification is a successful defense mechanism against adversarial attacks without requiring knowledge of the form of the incoming attack. Generally, adversarial purification aims to remove the adversarial perturbations therefore can make correct predictions based on the recovered clean samples. Despite the success of adversarial purification in the computer vision field that incorporates generative models such as energy-based models and diffusion models,using purification as a defense strategy against textual adversarial attacks is rarely explored. In this work, we introduce a novel adversarial purification method that focuses on defending against textual adversarial attacks. With the help of language models, we can inject noise by masking input texts and reconstructing the masked texts based on the masked language models. In this way, we construct an adversarial purification process for textual models against the most widely used word-substitution adversarial attacks. We test our proposed adversarial purification method on several strong adversarial attack methods including Textfooler and BERT-Attack and experimental results indicate that the purification algorithm can successfully defend against strong word-substitution attacks.
Discovering new intents is of great significance for establishing the Task-Oriented Dialogue System. Most existing methods either cannot transfer prior knowledge contained in known intents or fall into the dilemma of forgetting prior knowledge in the follow-up. Furthermore, these methods do not deeply explore the intrinsic structure of unlabeled data, and as a result, cannot seek out the characteristics that define an intent in general. In this paper, starting from the intuition that discovering intents could be beneficial for identifying known intents, we propose a probabilistic framework for discovering intents where intent assignments are treated as latent variables. We adopt the Expectation Maximization framework for optimization. Specifically, In the E-step, we conduct intent discovery and explore the intrinsic structure of unlabeled data by the posterior of intent assignments. In the M-step, we alleviate the forgetting of prior knowledge transferred from known intents by optimizing the discrimination of labeled data. Extensive experiments conducted on three challenging real-world datasets demonstrate the generality and effectiveness of the proposed framework and implementation.
Information Extraction (IE) spans several tasks with different output structures, such as named entity recognition, relation extraction and event extraction. Previously, those tasks were solved with different models because of diverse task output structures. Through re-examining IE tasks, we find that all of them can be interpreted as extracting spans and span relations. They can further be decomposed into token-pair classification tasks by using the start and end token of a span to pinpoint the span, and using the start-to-start and end-to-end token pairs of two spans to determine the relation. Based on the reformulation, we propose a Unified Token-pair Classification architecture for Information Extraction (UTC-IE), where we introduce Plusformer on top of the token-pair feature matrix. Specifically, it models axis-aware interaction with plus-shaped self-attention and local interaction with Convolutional Neural Network over token pairs. Experiments show that our approach outperforms task-specific and unified models on all tasks in 10 datasets, and achieves better or comparable results on 2 joint IE datasets. Moreover, UTC-IE speeds up over state-of-the-art models on IE tasks significantly in most datasets, which verifies the effectiveness of our architecture.
We present DiffusionBERT, a new generative masked language model based on discrete dif- fusion models. Diffusion models and many pre- trained language models have a shared training objective, i.e., denoising, making it possible to combine the two powerful models and enjoy the best of both worlds. On the one hand, dif- fusion models offer a promising training strat- egy that helps improve the generation quality. On the other hand, pre-trained denoising lan- guage models (e.g., BERT) can be used as a good initialization that accelerates convergence. We explore training BERT to learn the reverse process of a discrete diffusion process with an absorbing state and elucidate several designs to improve it. First, we propose a new noise schedule for the forward diffusion process that controls the degree of noise added at each step based on the information of each token. Sec- ond, we investigate several designs of incorpo- rating the time step into BERT. Experiments on unconditional text generation demonstrate that DiffusionBERT achieves significant improve- ment over existing diffusion models for text (e.g., D3PM and Diffusion-LM) and previous generative masked language models in terms of perplexity and BLEU score. Promising re- sults in conditional generation tasks show that DiffusionBERT can generate texts of compa- rable quality and more diverse than a series of established baselines.
In-context learning is a new learning paradigm where a language model conditions on a few input-output pairs (demonstrations) and a test input, and directly outputs the prediction. It has been shown sensitive to the provided demonstrations and thus promotes the research of demonstration retrieval: given a test input, relevant examples are retrieved from the training set to serve as informative demonstrations for in-context learning. While previous works train task-specific retrievers for several tasks separately, these methods are hard to transfer and scale on various tasks, and separately trained retrievers will cause a lot of parameter storage and deployment cost. In this paper, we propose Unified Demonstration Retriever (UDR), a single model to retrieve demonstrations for a wide range of tasks. To train UDR, we cast various tasks’ training signals into a unified list-wise ranking formulation by language model’s feedback. Then we propose a multi-task list-wise ranking training framework with an iterative mining strategy to find high-quality candidates, which can help UDR fully incorporate various tasks’ signals. Experiments on 30+ tasks across 13 task families and multiple data domains show that UDR significantly outperforms baselines. Further analyses show the effectiveness of each proposed component and UDR’s strong ability in various scenarios including different LMs (1.3B 175B), unseen datasets, varying demonstration quantities, etc. We will release the code and model checkpoint after review.
Discourse analysis is an important task because it models intrinsic semantic structures between sentences in a document. Discourse markers are natural representations of discourse in our daily language. One challenge is that the markers as well as pre-defined and human-labeled discourse relations can be ambiguous when describing the semantics between sentences. We believe that a better approach is to use a contextual-dependent distribution over the markers to express discourse information. In this work, we propose to learn a Distributed Marker Representation (DMR) by utilizing the (potentially) unlimited discourse marker data with a latent discourse sense, thereby bridging markers with sentence pairs. Such representations can be learned automatically from data without supervision, and in turn provide insights into the data itself. Experiments show the SOTA performance of our DMR on the implicit discourse relation recognition task and strong interpretability. Our method also offers a valuable tool to understand complex ambiguity and entanglement among discourse markers and manually defined discourse relations.
Out-of-domain (OOD) intent classification is an active field of natural language understanding, which is of great practical significance for intelligent devices such as the Task-Oriented Dialogue System. It mainly contains two challenges: it requires the model to know what it knows and what it does not know. This paper investigates “overthinking” in the open-world scenario and its impact on OOD intent classification. Inspired by this, we propose a two-birds-one-stone method, which allows the model to decide whether to make a decision on OOD classification early during inference and can ensure accuracy and accelerate inference. At the same time, to adapt to the behavior of dynamic inference, we also propose a training method based on ensemble methods. In addition to bringing certain theoretical insights, we also conduct detailed experiments on three real-world intent datasets. Compared with the previous baselines, our method can not only improve inference speed, but also achieve significant performance improvements.
Prompt tuning is a parameter-efficient approach to adapting pre-trained language models to downstream tasks. Although prompt tuning has been shown to match the performance of full model tuning when training data is sufficient, it tends to struggle in few-shot learning settings. In this paper, we present Multi-task Pre-trained Modular Prompt (MP2) to boost prompt tuning for few-shot learning. MP2 is a set of combinable prompts pre-trained on 38 Chinese tasks. On downstream tasks, the pre-trained prompts are selectively activated and combined, leading to strong compositional generalization to unseen tasks. To bridge the gap between pre-training and fine-tuning, we formulate upstream and downstream tasks into a unified machine reading comprehension task. Extensive experiments under two learning paradigms, i.e., gradient descent and black-box tuning, show that MP2 significantly outperforms prompt tuning, full model tuning, and prior prompt pre-training methods in few-shot settings. In addition, we demonstrate that MP2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.
Recent works have introduced Abstract Meaning Representation (AMR) for Document-level Event Argument Extraction (Doc-level EAE), since AMR provides a useful interpretation of complex semantic structures and helps to capture long-distance dependency. However, in these works AMR is used only implicitly, for instance, as additional features or training signals. Motivated by the fact that all event structures can be inferred from AMR, this work reformulates EAE as a link prediction problem on AMR graphs. Since AMR is a generic structure and does not perfectly suit EAE, we propose a novel graph structure, Tailored AMR Graph (TAG), which compresses less informative subgraphs and edge types, integrates span information, and highlights surrounding events in the same document. With TAG, we further propose a novel method using graph neural networks as a link prediction model to find event arguments. Our extensive experiments on WikiEvents and RAMS show that this simpler approach outperforms the state-of-the-art models by 3.63pt and 2.33pt F1, respectively, and do so with reduced 56% inference time.
Recent works show the effectiveness of cache-based neural coreference resolution models on long documents. These models incrementally process a long document from left to right and extract relations between mentions and entities in a cache, resulting in much lower memory and computation cost compared to computing all mentions in parallel. However, they do not handle cache misses when high-quality entities are purged from the cache, which causes wrong assignments and leads to prediction errors. We propose a new hybrid cache that integrates two eviction policies to capture global and local entities separately, and effectively reduces the aggregated cache misses up to half as before, while improving F1 score of coreference by 0.7 5.7pt. As such, the hybrid policy can accelerate existing cache-based models and offer a new long document coreference resolution solution. Results show that our method outperforms existing methods on four benchmarks while saving up to 83% of inference time against non-cache-based models. Further, we achieve a new state-of-the-art on a long document coreference benchmark, LitBank.
Large language models (LLMs) pre-trained on massive corpora have demonstrated impressive few-shot learning ability on many NLP tasks. A common practice is to recast the task into a text-to-text format such that generative LLMs of natural language (NL-LLMs) like GPT-3 can be prompted to solve it. However, it is nontrivial to perform information extraction (IE) tasks with NL-LLMs since the output of the IE task is usually structured and therefore is hard to be converted into plain text. In this paper, we propose to recast the structured output in the form of code instead of natural language and utilize generative LLMs of code (Code-LLMs) such as Codex to perform IE tasks, in particular, named entity recognition and relation extraction. In contrast to NL-LLMs, we show that Code-LLMs can be well-aligned with these IE tasks by designing code-style prompts and formulating these IE tasks as code generation tasks. Experiment results on seven benchmarks show that our method consistently outperforms fine-tuning moderate-size pre-trained models specially designed for IE tasks (e.g., UIE) and prompting NL-LLMs under few-shot settings. We further conduct a series of in-depth analyses to demonstrate the merits of leveraging Code-LLMs for IE tasks.
Named entity recognition (NER) is the task to detect and classify entity spans in the text. When entity spans overlap between each other, the task is named as nested NER. Span-based methods have been widely used to tackle nested NER. Most of these methods get a score matrix, where each entry corresponds to a span. However, previous work ignores spatial relations in the score matrix. In this paper, we propose using Convolutional Neural Network (CNN) to model these spatial relations. Despite being simple, experiments in three commonly used nested NER datasets show that our model surpasses several recently proposed methods with the same pre-trained encoders. Further analysis shows that using CNN can help the model find more nested entities. Besides, we find that different papers use different sentence tokenizations for the three nested NER datasets, which will influence the comparison. Thus, we release a pre-processing script to facilitate future comparison.
While pre-trained Chinese language models have demonstrated impressive performance on a wide range of NLP tasks, the Chinese Spell Checking (CSC) task remains a challenge. Previous research has explored using information such as glyphs and phonetics to improve the ability of CSC models to distinguish misspelled characters, with good results at the accuracy level on public datasets. However, the generalization ability of these CSC models has not been well understood: it is unclear whether they incorporate glyph-phonetic information and, if so, whether this information is fully utilized. In this paper, we aim to better understand the role of glyph-phonetic information in the CSC task and suggest directions for improvement. Additionally, we propose a new, more challenging, and practical setting for testing the generalizability of CSC models. All code is made publicly available.
Out-of-Domain (OOD) Intent Classification and New Intent Discovering are two basic and critical tasks in the Task-Oriented Dialogue System, which are typically treated two independent tasks. Classification focuses on identifying intents beyond the predefined set of the dialog system, but it will not further differentiate detected OOD intents in fine granularity. Discovering focuses on how to cluster unlabeled samples according to their semantic representation, which relies heavily on prior knowledge and can not provide label information for the formed clusters. To be closer to the real user-facing scenarios, we introduce a task paradigm to extend Classification with Discovering referred as Open Environment Intent Prediction, which is to make a further fine-grained discovery of OOD based on OOD Intent Classification. Using various widely-used generative models as an archetype, we propose a general scheme for Open Environment Intent Prediction. In a nutshell, we first perform intent detection to identify the In-domain (IND) samples and then generate labels for those identified as OOD. With these generated labels, we can discover new general intents and provide label information for them. We develop a suite of benchmarks on the existing intent datasets and present a simple yet effective implementation. Extensive experiments demonstrate that our method establishes substantial improvement compared to the baselines.
Large language models (LLMs) have a wealth of knowledge that allows them to excel in various Natural Language Processing (NLP) tasks. Current research focuses on enhancing their performance within their existing knowledge. Despite their vast knowledge, LLMs are still limited by the amount of information they can accommodate and comprehend. Therefore, the ability to understand their own limitations on the unknows, referred to as self-knowledge, is of paramount importance. This study aims to evaluate LLMs’ self-knowledge by assessing their ability to identify unanswerable or unknowable questions. We introduce an automated methodology to detect uncertainty in the responses of these models, providing a novel measure of their self-knowledge. We further introduce a unique dataset, SelfAware, consisting of unanswerable questions from five diverse categories and their answerable counterparts. Our extensive analysis, involving 20 LLMs including GPT-3, InstructGPT, and LLaMA, discovering an intrinsic capacity for self-knowledge within these models. Moreover, we demonstrate that in-context learning and instruction tuning can further enhance this self-knowledge. Despite this promising insight, our findings also highlight a considerable gap between the capabilities of these models and human proficiency in recognizing the limits of their knowledge.
Dialogue data in real scenarios tend to be sparsely available, rendering data-starved end-to-end dialogue systems trained inadequately. We discover that data utilization efficiency in low-resource scenarios can be enhanced by mining alignment information uncertain utterance and deterministic dialogue state. Therefore, we innovatively implement dual learning in task-oriented dialogues to exploit the correlation of heterogeneous data. In addition, the one-to-one duality is converted into a multijugate duality to reduce the influence of spurious correlations in dual training for generalization. Without introducing additional parameters, our method could be implemented in arbitrary networks. Extensive empirical analyses demonstrate that our proposed method improves the effectiveness of end-to-end task-oriented dialogue systems under multiple benchmarks and obtains state-of-the-art results in low-resource scenarios.
Contrastive learning has become a popular approach in natural language processing, particularly for the learning of sentence embeddings.However, the discrete nature of natural language makes it difficult to ensure the quality of positive and negative sample pairs generated through data augmentation methods. Although supervised contrastive learning can produce more accurate sample pairs with human feedback labels, it still lacks fine-grained training signals. In this paper, we propose to improve Contrastive Learning of sentence embeddings from AI Feedback (CLAIF).Our method utilizes AI feedback from large pre-trained language models (LLMs) to construct sample pairs with fine-grained sample similarity scores to improve contrastive learning. Besides, we combine human feedback and AI feedback to provide better supervision signals for supervised contrastive learning of sentence embeddings.Experimental results show that our method achieves state-of-the-art performance on several semantic textual similarity (STS) and transfer learning tasks compared to other unsupervised and supervised contrastive learning methods.
Abuse of large language models reveals high risks as large language models are being deployed at an astonishing speed. It is important to protect the model weights to avoid malicious usage that violates licenses of open-source large language models. This paper proposes a novel watermarking strategy that plants watermarks in the quantization process of large language models without pre-defined triggers during inference. The watermark works when the model is used in the fp32 mode and remains hidden when the model is quantized to int8, in this way, the users can only inference the model without further supervised fine-tuning of the model. We successfully plant the watermark into open-source large language model weights including GPT-Neo and LLaMA. We hope our proposed method can provide a potential direction for protecting model weights in the era of large language model applications.
In-context learning is a new learning paradigm where a language model observes a few examples and directly outputs the test input’s prediction. Previous works have shown that it is sensitive to the provided examples and randomly sampled examples probably cause inferior performance. In this paper, we propose finding “support examples” for in-context learning: Given a training dataset, it aims to select one permutation of a few examples, which can well characterize the task for in-context learning and thus lead to superior performance. Although for traditional gradient-based training, there are extensive methods to find a coreset from the entire dataset, they struggle to find important in-context examples, because in-context learning occurs in the language model’s forward process without gradients or parameter updates and thus has a significant gap with traditional training. Additionally, the strong dependence among in-context examples makes it an NP-hard combinatorial optimization problem and enumerating all permutations is infeasible. Hence we propose **LENS**, a fi**L**ter-th**EN**-**S**earch method to tackle this challenge in two stages: irst we filter the dataset to obtain individually informative in-context examples. Specifically, we propose a novel metric, InfoScore, to evaluate the example’s in-context informativeness based on the language model’s feedback, and further propose a progressive filtering process to filter out uninformative examples. Then we propose diversity-guided example search which iteratively refines and evaluates the selected example permutations, to find examples that fully depict the task. The experimental results show that LENS significantly outperforms a wide range of baselines and further analyses show that each component contribute critically to the improvements and shed light on the principles of supporting examples and in-context learning.
With the rapid development of neural network applications in NLP, model robustness problem is gaining more attention. Different from computer vision, the discrete nature of texts makes it more challenging to explore robustness in NLP. Therefore, in this paper, we aim to connect discrete perturbations with continuous perturbations, therefore we can use such connections as a bridge to help understand discrete perturbations in NLP models. Specifically, we first explore how to connect and measure the correlation between discrete perturbations and continuous perturbations. Then we design a regression task as a PerturbScore to learn the correlation automatically. Through experimental results, we find that we can build a connection between discrete and continuous perturbations and use the proposed PerturbScore to learn such correlation, surpassing previous methods used in discrete perturbation measuring. Further, the proposed PerturbScore can be well generalized to different datasets, perturbation methods, indicating that we can use it as a powerful tool to study model robustness in NLP.
Multi-modal large language models are regarded as a crucial step towards Artificial General Intelligence (AGI) and have garnered significant interest with the emergence of ChatGPT. However, current speech-language models typically adopt the cascade paradigm, preventing inter-modal knowledge transfer. In this paper, we propose SpeechGPT, a large language model with intrinsic cross-modal conversational abilities, capable of perceiving and generating multi-modal content. With discrete speech representations, we construct SpeechInstruct, the first large-scale cross-modal speech instruction dataset. Additionally, we employ a three-stage training strategy that includes modality-adaptation pre-training, cross-modal instruction fine-tuning, and chain-of-modality instruction fine-tuning. The experimental results demonstrate that SpeechGPT has an impressive capacity to follow cross-modal human instructions and highlight the potential of handling multiple modalities with one model. Code and models are available in https://github.com/0nutation/SpeechGPT. Demos are shown in https://0nutation.github.io/SpeechGPT.github.io/.
Widely applied large language models (LLMs) can generate human-like content, raising concerns about the abuse of LLMs. Therefore, it is important to build strong AI-generated text (AIGT) detectors. Current works only consider document-level AIGT detection, therefore, in this paper, we first introduce a sentence-level detection challenge by synthesizing a dataset that contains documents that are polished with LLMs, that is, the documents contain sentences written by humans and sentences modified by LLMs. Then we propose Sequence X (Check) GPT, a novel method that utilizes log probability lists from white-box LLMs as features for sentence-level AIGT detection. These features are composed like waves in speech processing and cannot be studied by LLMs. Therefore, we build SeqXGPT based on convolution and self-attention networks. We test it in both sentence and document-level detection challenges. Experimental results show that previous methods struggle in solving sentence-level AIGT detection, while our method not only significantly surpasses baseline methods in both sentence and document-level detection challenges but also exhibits strong generalization capabilities.
As large language models (LLMs) have shown effectiveness with different prompting methods, such as Chain of Thought, Program of Thought, we find that these methods have formed a great complementarity to each other on math reasoning tasks. In this work, we propose XoT, an integrated problem solving framework by prompting LLMs with diverse reasoning thoughts. For each question, XoT always begins with selecting the most suitable method then executes each method iteratively. Within each iteration, XoT actively checks the validity of the generated answer and incorporates the feedback from external executors, allowing it to dynamically switch among different prompting methods. Through extensive experiments on 10 popular math reasoning datasets, we demonstrate the effectiveness of our proposed approach and thoroughly analyze the strengths of each module. Moreover, empirical results suggest that our framework is orthogonal to recent work that makes improvements on single reasoning methods and can further generalise to logical reasoning domain. By allowing method switching, XoT provides a fresh perspective on the collaborative integration of diverse reasoning thoughts in a unified framework.
Large Language Models (LLMs) have shown impressive abilities on various tasks. However, fundamentally improving them depends on high-quality datasets or computationally expensive fine-tuning. On the contrary, humans can easily improve themselves by self-thinking and memory, without external resources. In this paper, we propose a framework, **MoT**, to let the LLM self-improve through **M**emory **o**f **T**houghts, without annotated datasets and parameter updates. Specifically, MoT is divided into two stages: 1. before the test stage, the LLM pre-thinks on the unlabeled dataset and saves the high-confidence thoughts as external memory; 2. During the test stage, given a test question, the LLM recalls relevant memory to help itself reason and answer it. Experimental results show that MoT can help ChatGPT significantly improve its abilities in arithmetic reasoning, commonsense reasoning, factual reasoning, and natural language inference. Further analyses show that each component contributes critically to the improvements and MoT can lead to consistent improvements across various CoT methods and LLMs.
Large language models (LLMs) can be used to serve as agents to simulate human behaviors, given the powerful ability to understand human instructions and provide high-quality generated texts. Such ability stimulates us to wonder whether LLMs can simulate a person in a higher form than simple human behaviors. Therefore, we aim to train an agent with the profile, experience, and emotional states of a specific person instead of using limited prompts to instruct ChatGPT API. In this work, we introduce Character-LLM that teach LLMs to act as specific people such as Beethoven, Queen Cleopatra, Julius Caesar, etc. Our method focuses on editing profiles as experiences of a certain character and training models to be personal simulacra with these experiences. To assess the effectiveness of our approach, we build a test playground that interviews trained agents and evaluates whether the agents memorize their characters and experiences. Experimental results show interesting observations that help build future simulacra of humankind.
Large Language Models (LLMs) have recently made significant strides in complex reasoning tasks through the Chain-of-Thought technique. Despite this progress, their reasoning is often constrained by their intrinsic understanding, lacking external insights. To address this, we propose Exchange-of-Thought (EoT), a novel framework that enables cross-model communication during problem-solving. Drawing inspiration from network topology, EoT integrates four unique communication paradigms: Memory, Report, Relay, and Debate. This paper delves into the communication dynamics and volume associated with each paradigm. To counterbalance the risks of incorrect reasoning chains, we implement a robust confidence evaluation mechanism within these communications. Our experiments across diverse complex reasoning tasks demonstrate that EoT significantly surpasses established baselines, underscoring the value of external insights in enhancing LLM performance. Furthermore, we show that EoT achieves these superior results in a cost-effective manner, marking a promising advancement for efficient and collaborative AI problem-solving.
Large language models (LLMs) are increasingly pivotal in a wide range of natural language processing tasks. Access to pre-trained models, courtesy of the open-source community, has made it possible to adapt these models to specific applications for enhanced performance. However, the substantial resources required for training these models necessitate efficient solutions. This paper introduces CoLLiE, an efficient library that facilitates collaborative training of large language models using 3D parallelism, parameter-efficient fine-tuning (PEFT) methods, and optimizers such as Lion, Adan, Sophia, and LOMO. With its modular design and comprehensive functionality, CoLLiE offers a balanced blend of efficiency, ease of use, and customization. CoLLiE has proven superior training efficiency in comparison with prevalent solutions in pre-training and fine-tuning scenarios. Furthermore, we provide an empirical evaluation of the correlation between model size and GPU memory consumption under different optimization methods, as well as an analysis of the throughput. Lastly, we carry out a comprehensive comparison of various optimizers and PEFT methods within the instruction-tuning context. CoLLiE is available at https://github.com/OpenLMLab/collie.
Supersized pre-trained language models have pushed the accuracy of various natural language processing (NLP) tasks to a new state-of-the-art (SOTA). Rather than pursuing the reachless SOTA accuracy, more and more researchers start paying attention to model efficiency and usability. Different from accuracy, the metric for efficiency varies across different studies, making them hard to be fairly compared. To that end, this work presents ELUE (Efficient Language Understanding Evaluation), a standard evaluation, and a public leaderboard for efficient NLP models. ELUE is dedicated to depicting the Pareto Frontier for various language understanding tasks, such that it can tell whether and how much a method achieves Pareto improvement. Along with the benchmark, we also release a strong baseline, ElasticBERT, which allows BERT to exit at any layer in both static and dynamic ways. We demonstrate the ElasticBERT, despite its simplicity, outperforms or performs on par with SOTA compressed and early exiting models. With ElasticBERT, the proposed ELUE has a strong Pareto Frontier and makes a better evaluation for efficient NLP models.
The Out-of-Domain (OOD) intent classification is a basic and challenging task for dialogue systems. Previous methods commonly restrict the region (in feature space) of In-domain (IND) intent features to be compact or simply-connected implicitly, which assumes no OOD intents reside, to learn discriminative semantic features. Then the distribution of the IND intent features is often assumed to obey a hypothetical distribution (Gaussian mostly) and samples outside this distribution are regarded as OOD samples. In this paper, we start from the nature of OOD intent classification and explore its optimization objective. We further propose a simple yet effective method, named KNN-contrastive learning. Our approach utilizes k-nearest neighbors (KNN) of IND intents to learn discriminative semantic features that are more conducive to OOD detection. Notably, the density-based novelty detection algorithm is so well-grounded in the essence of our method that it is reasonable to use it as the OOD detection algorithm without making any requirements for the feature distribution. Extensive experiments on four public datasets show that our approach can not only enhance the OOD detection performance substantially but also improve the IND intent classification while requiring no restrictions on feature distribution.
In this paper, we propose the CodeRetriever model, which learns the function-level code semantic representations through large-scale code-text contrastive pre-training. We adopt two contrastive learning schemes in CodeRetriever: unimodal contrastive learning and bimodal contrastive learning. For unimodal contrastive learning, we design an unsupervised learning approach to build semantic-related code pairs based on the documentation and function name. For bimodal contrastive learning, we leverage the documentation and in-line comments of code to build code-text pairs. Both contrastive objectives can fully leverage large-scale code corpus for pre-training. Extensive experimental results show that CodeRetriever achieves new state-of-the-art with significant improvement over existing code pre-trained models, on eleven domain/language-specific code search tasks with six programming languages in different code granularity (function-level, snippet-level and statement-level).These results demonstrate the effectiveness and robustness of CodeRetriever.The codes and resources are available at https://github.com/microsoft/AR2/tree/main/CodeRetriever.
Automatic evaluation metrics are crucial to the development of generative systems. In recent years, pre-trained language model (PLM) based metrics, such as BERTScore, have been commonly adopted in various generation tasks. However, it has been demonstrated that PLMs encode a range of stereotypical societal biases, leading to a concern about the fairness of PLMs as metrics. To that end, this work presents the first systematic study on the social bias in PLM-based metrics. We demonstrate that popular PLM-based metrics exhibit significantly higher social bias than traditional metrics on 6 sensitive attributes, namely race, gender, religion, physical appearance, age, and socioeconomic status. In-depth analysis suggests that choosing paradigms (matching, regression, or generation) of the metric has a greater impact on fairness than choosing PLMs. In addition, we develop debiasing adapters that are injected into PLM layers, mitigating bias in PLM-based metrics while retaining high performance for evaluating text generation.
Most downstream adaptation methods tune all or part of the parameters of pre-trained models (PTMs) through gradient descent, where the tuning cost increases linearly with the growth of the model size.By contrast, gradient-free methods only require the forward computation of the PTM to tune the prompt, retaining the benefits of efficient tuning and deployment.Though, past work on gradient-free tuning often introduces gradient descent to seek a good initialization of prompt and lacks versatility across tasks and PTMs.In this paper, we present BBTv2, an improved version of Black-Box Tuning, to drive PTMs for few-shot learning.We prepend continuous prompts to every layer of the PTM and propose a divide-and-conquer gradient-free algorithm to optimize the prompts at different layers alternately.Extensive experiments across various tasks and PTMs show that BBTv2 can achieve comparable performance to full model tuning and state-of-the-art parameter-efficient methods (e.g., Adapter, LoRA, BitFit, etc.) under few-shot settings while maintaining much fewer tunable parameters.
Interpreting the reasoning process from questions to answers poses a challenge in approaching explainable QA. A recently proposed structured reasoning format, entailment tree, manages to offer explicit logical deductions with entailment steps in a tree structure. To generate entailment trees, prior single pass sequence-to-sequence models lack visible internal decision probability, while stepwise approaches are supervised with extracted single step data and cannot model the tree as a whole. In this work, we propose RLET, a Reinforcement Learning based Entailment Tree generation framework, which is trained utilising the cumulative signals across the whole tree. RLET iteratively performs single step reasoning with sentence selection and deduction generation modules, from which the training signal is accumulated across the tree with elaborately designed aligned reward function that is consistent with the evaluation. To the best of our knowledge, we are the first to introduce RL into the entailment tree generation task. Experiments on three settings of the EntailmentBank dataset demonstrate the strength of using RL framework.
Whole word masking (WWM), which masks all subwords corresponding to a word at once, makes a better English BERT model. For the Chinese language, however, there is no subword because each token is an atomic character. The meaning of a word in Chinese is different in that a word is a compositional unit consisting of multiple characters. Such difference motivates us to investigate whether WWM leads to better context understanding ability for Chinese BERT. To achieve this, we introduce two probing tasks related to grammatical error correction and ask pretrained models to revise or insert tokens in a masked language modeling manner. We construct a dataset including labels for 19,075 tokens in 10,448 sentences. We train three Chinese BERT models with standard character-level masking (CLM), WWM, and a combination of CLM and WWM, respectively. Our major findings are as follows: First, when one character needs to be inserted or replaced, the model trained with CLM performs the best. Second, when more than one character needs to be handled, WWM is the key to better performance. Finally, when being fine-tuned on sentence-level downstream tasks, models trained with different masking strategies perform comparably.
Early exiting allows instances to exit at different layers according to the estimation of difficulty. Previous works usually adopt heuristic metrics such as the entropy of internal outputs to measure instance difficulty, which suffers from generalization and threshold-tuning. In contrast, learning to exit, or learning to predict instance difficulty is a more appealing way. Though some effort has been devoted to employing such “learn-to-exit” modules, it is still unknown whether and how well the instance difficulty can be learned. As a response, we first conduct experiments on the learnability of instance difficulty, which demonstrates that modern neural models perform poorly on predicting instance difficulty. Based on this observation, we propose a simple-yet-effective Hash-based Early Exiting approach HashEE) that replaces the learn-to-exit modules with hash functions to assign each token to a fixed exiting layer. Different from previous methods, HashEE requires no internal classifiers nor extra parameters, and therefore is more efficient. HashEE can be used in various tasks (including language understanding and generation) and model architectures such as seq2seq models. Experimental results on classification, regression, and generation tasks demonstrate that HashEE can achieve higher performance with fewer FLOPs and inference time compared with previous state-of-the-art early exiting methods.
Code contrastive pre-training has recently achieved significant progress on code-related tasks. In this paper, we present SCodeR, a Soft-labeled contrastive pre-training framework with two positive sample construction methods to learn functional-level CodeRepresentation. Considering the relevance between codes in a large-scale code corpus, the soft-labeled contrastive pre-training can obtain fine-grained soft-labels through an iterative adversarial manner and use them to learn better code representation. The positive sample construction is another key for contrastive pre-training. Previous works use transformation-based methods like variable renaming to generate semantically equal positive codes. However, they usually result in the generated code with a highly similar surface form, and thus mislead the model to focus on superficial code structure instead of code semantics. To encourage SCodeR to capture semantic information from the code, we utilize code comments and abstract syntax sub-trees of the code to build positive samples. We conduct experiments on four code-related tasks over seven datasets. Extensive experimental results show that SCodeR achieves new state-of-the-art performance on all of them, which illustrates the effectiveness of the proposed pre-training method.
Dialogue meaning representation formulates natural language utterance semantics in their conversational context in an explicit and machine-readable form. Previous work typically follows the intent-slot framework, which is easy for annotation yet limited in scalability for complex linguistic expressions. A line of works alleviates the representation issue by introducing hierarchical structures but challenging to express complex compositional semantics, such as negation and coreference. We propose Dialogue Meaning Representation (DMR), a pliable and easily extendable representation for task-oriented dialogue. Our representation contains a set of nodes and edges to represent rich compositional semantics. Moreover, we propose an inheritance hierarchy mechanism focusing on domain extensibility. Additionally, we annotated DMR-FastFood, a multi-turn dialogue dataset with more than 70k utterances, with DMR. We propose two evaluation tasks to evaluate different dialogue models and a novel coreference resolution model GNNCoref for the graph-based coreference resolution task. Experiments show that DMR can be parsed well with pre-trained Seq2Seq models, and GNNCoref outperforms the baseline models by a large margin.The dataset and code are available at https://github.com/amazon-research/dialogue-meaning-representation
Task-Oriented Dialogue (TOD) systems are drawing more and more attention in recent studies.Current methods focus on constructing pre-trained models or fine-tuning strategies while the evaluation of TOD is limited by a policy mismatch problem.That is, during evaluation, the user utterances are from the annotated dataset while these utterances should interact with previous responses which can have many alternatives besides annotated texts.Therefore, in this work, we propose an interactive evaluation framework for TOD. We first build a goal-oriented user simulator based on pre-trained models and then use the user simulator to interact with the dialogue system to generate dialogues.Besides, we introduce a sentence-level and a session-level score to measure the sentence fluency and session coherence in the interactive evaluation. Experimental results show that RL-based TOD systems trained by our proposed user simulator can achieve nearly 98% inform and success rates in the interactive evaluation of MultiWOZ dataset and the proposed scores measure the response quality besides the inform and success rates.We are hoping that our work will encourage simulator-based interactive evaluations in the TOD task.
Prompt tuning is a parameter-efficient tuning (PETuning) method for utilizing pre-trained models (PTMs) that simply prepends a soft prompt to the input and only optimizes the prompt to adapt PTMs to downstream tasks. Although it is parameter- and deployment-efficient, its performance still lags behind other state-of-the-art PETuning methods. Besides, the training cost of prompt tuning is not significantly reduced due to the back-propagation through the entire model. Through empirical analyses, we shed some light on the lagging performance of prompt tuning and recognize a trade-off between the propagation distance from label signals to the inserted prompt and the influence of the prompt on model outputs. Further, we present Late Prompt Tuning (LPT) that inserts a late prompt into an intermediate layer of the PTM instead of the input layer or all layers. The late prompt is obtained by a neural prompt generator conditioned on the hidden states before the prompt insertion layer and therefore is instance-dependent. Through extensive experimental results across various tasks and PTMs, we show that LPT can achieve competitive performance to full model tuning and other PETuning methods under both full-data and few-shot scenarios while possessing faster training speed and lower memory cost.
In recent years, there is a surge of generation-based information extraction work, which allows a more direct use of pre-trained language models and efficiently captures output dependencies. However, previous generative methods using lexical representation do not naturally fit document-level relation extraction (DocRE) where there are multiple entities and relational facts. In this paper, we investigate the root cause of the underwhelming performance of the existing generative DocRE models and discover that the culprit is the inadequacy of the training paradigm, instead of the capacities of the models. We propose to generate a symbolic and ordered sequence from the relation matrix which is deterministic and easier for model to learn. Moreover, we design a parallel row generation method to process overlong target sequences. Besides, we introduce several negative sampling strategies to improve the performance with balanced signals. Experimental results on four datasets show that our proposed method can improve the performance of the generative DocRE models.
Generalized text representations are the foundation of many natural language understanding tasks. To fully utilize the different corpus, it is inevitable that models need to understand the relevance among them. However, many methods ignore the relevance and adopt a single-channel model (a coarse paradigm) directly for all tasks, which lacks enough rationality and interpretation. In addition, some existing works learn downstream tasks by stitches skill block (a fine paradigm), which might cause irrational results due to its redundancy and noise. In this work, we first analyze the task correlation through three different perspectives, , data property, manual design, and model-based relevance, based on which the similar tasks are grouped together. Then, we propose a hierarchical framework with a coarse-to-fine paradigm, with the bottom level shared to all the tasks, the mid-level divided to different groups, and the top-level assigned to each of the tasks. This allows our model to learn basic language properties from all tasks, boost performance on relevant tasks, and reduce the negative impact from irrelevant tasks. Our experiments on 13 benchmark datasets across five natural language understanding tasks demonstrate the superiority of our method.
Traditional training paradigms for extractive and abstractive summarization systems always only use token-level or sentence-level training objectives. However, the output summary is always evaluated from summary-level which leads to the inconsistency in training and evaluation. In this paper, we propose a Contrastive Learning based re-ranking framework for one-stage summarization called CoLo. By modeling a contrastive objective, we show that the summarization model is able to directly generate summaries according to the summary-level score without additional modules and parameters. Extensive experiments demonstrate that CoLo boosts the extractive and abstractive results of one-stage systems on CNN/DailyMail benchmark to 44.58 and 46.33 ROUGE-1 score while preserving the parameter efficiency and inference efficiency. Compared with state-of-the-art multi-stage systems, we save more than 100 GPU training hours and obtaining 3x 8x speed-up ratio during inference while maintaining comparable results.
Pre-trained models have brought remarkable success on the text summarization task. For dialogue summarization, the subdomain of text summarization, utterances are concatenated to flat text before being processed. As a result, existing summarization systems based on pre-trained models are unable to recognize the unique format of the speaker-utterance pair well in the dialogue. To investigate this issue, we conduct probing tests and manual analysis, and find that the powerful pre-trained model can not identify different speakers well in the conversation, which leads to various factual errors. Moreover, we propose three speaker-aware supervised contrastive learning (SCL) tasks: Token-level SCL, Turn-level SCL, and Global-level SCL. Comprehensive experiments demonstrate that our methods achieve significant performance improvement on two mainstream dialogue summarization datasets. According to detailed human evaluations, pre-trained models equipped with SCL tasks effectively generate summaries with better factual consistency.
Both performance and efficiency are crucial factors for sequence labeling tasks in many real-world scenarios. Although the pre-trained models (PTMs) have significantly improved the performance of various sequence labeling tasks, their computational cost is expensive. To alleviate this problem, we extend the recent successful early-exit mechanism to accelerate the inference of PTMs for sequence labeling tasks. However, existing early-exit mechanisms are specifically designed for sequence-level tasks, rather than sequence labeling. In this paper, we first propose a simple extension of sentence-level early-exit for sequence labeling tasks. To further reduce the computational cost, we also propose a token-level early-exit mechanism that allows partial tokens to exit early at different layers. Considering the local dependency inherent in sequence labeling, we employed a window-based criterion to decide for a token whether or not to exit. The token-level early-exit brings the gap between training and inference, so we introduce an extra self-sampling fine-tuning stage to alleviate it. The extensive experiments on three popular sequence labeling tasks show that our approach can save up to 66%∼75% inference cost with minimal performance degradation. Compared with competitive compressed models such as DistilBERT, our approach can achieve better performance under the same speed-up ratios of 2×, 3×, and 4×.
Aspect-based Sentiment Analysis (ABSA) aims to identify the aspect terms, their corresponding sentiment polarities, and the opinion terms. There exist seven subtasks in ABSA. Most studies only focus on the subsets of these subtasks, which leads to various complicated ABSA models while hard to solve these subtasks in a unified framework. In this paper, we redefine every subtask target as a sequence mixed by pointer indexes and sentiment class indexes, which converts all ABSA subtasks into a unified generative formulation. Based on the unified formulation, we exploit the pre-training sequence-to-sequence model BART to solve all ABSA subtasks in an end-to-end framework. Extensive experiments on four ABSA datasets for seven subtasks demonstrate that our framework achieves substantial performance gain and provides a real unified end-to-end solution for the whole ABSA subtasks, which could benefit multiple tasks.
Named Entity Recognition (NER) is the task of identifying spans that represent entities in sentences. Whether the entity spans are nested or discontinuous, the NER task can be categorized into the flat NER, nested NER, and discontinuous NER subtasks. These subtasks have been mainly solved by the token-level sequence labelling or span-level classification. However, these solutions can hardly tackle the three kinds of NER subtasks concurrently. To that end, we propose to formulate the NER subtasks as an entity span sequence generation task, which can be solved by a unified sequence-to-sequence (Seq2Seq) framework. Based on our unified framework, we can leverage the pre-trained Seq2Seq model to solve all three kinds of NER subtasks without the special design of the tagging schema or ways to enumerate spans. We exploit three types of entity representations to linearize entities into a sequence. Our proposed framework is easy-to-implement and achieves state-of-the-art (SoTA) or near SoTA performance on eight English NER datasets, including two flat NER datasets, three nested NER datasets, and three discontinuous NER datasets.
We present fastHan, an open-source toolkit for four basic tasks in Chinese natural language processing: Chinese word segmentation (CWS), Part-of-Speech (POS) tagging, named entity recognition (NER), and dependency parsing. The backbone of fastHan is a multi-task model based on a pruned BERT, which uses the first 8 layers in BERT. We also provide a 4-layer base model compressed from the 8-layer model. The joint-model is trained and evaluated on 13 corpora of four tasks, yielding near state-of-the-art (SOTA) performance in dependency parsing and NER, achieving SOTA performance in CWS and POS. Besides, fastHan’s transferability is also strong, performing much better than popular segmentation tools on a non-training corpus. To better meet the need of practical application, we allow users to use their own labeled data to further fine-tune fastHan. In addition to its small size and excellent performance, fastHan is user-friendly. Implemented as a python package, fastHan isolates users from the internal technical details and is convenient to use. The project is released on Github.
TextFlint is a multilingual robustness evaluation toolkit for NLP tasks that incorporates universal text transformation, task-specific transformation, adversarial attack, subpopulation, and their combinations to provide comprehensive robustness analyses. This enables practitioners to automatically evaluate their models from various aspects or to customize their evaluations as desired with just a few lines of code. TextFlint also generates complete analytical reports as well as targeted augmented data to address the shortcomings of the model in terms of its robustness. To guarantee acceptability, all the text transformations are linguistically based and all the transformed data selected (up to 100,000 texts) scored highly under human evaluation. To validate the utility, we performed large-scale empirical evaluations (over 67,000) on state-of-the-art deep learning models, classic supervised methods, and real-world systems. The toolkit is already available at https://github.com/textflint with all the evaluation results demonstrated at textflint.io.
Aspect-based Sentiment Analysis (ABSA), aiming at predicting the polarities for aspects, is a fine-grained task in the field of sentiment analysis. Previous work showed syntactic information, e.g. dependency trees, can effectively improve the ABSA performance. Recently, pre-trained models (PTMs) also have shown their effectiveness on ABSA. Therefore, the question naturally arises whether PTMs contain sufficient syntactic information for ABSA so that we can obtain a good ABSA model only based on PTMs. In this paper, we firstly compare the induced trees from PTMs and the dependency parsing trees on several popular models for the ABSA task, showing that the induced tree from fine-tuned RoBERTa (FT-RoBERTa) outperforms the parser-provided tree. The further analysis experiments reveal that the FT-RoBERTa Induced Tree is more sentiment-word-oriented and could benefit the ABSA task. The experiments also show that the pure RoBERTa-based model can outperform or approximate to the previous SOTA performances on six datasets across four languages since it implicitly incorporates the task-oriented syntactic information.
Recent researches show that pre-trained models (PTMs) are beneficial to Chinese Word Segmentation (CWS). However, PTMs used in previous works usually adopt language modeling as pre-training tasks, lacking task-specific prior segmentation knowledge and ignoring the discrepancy between pre-training tasks and downstream CWS tasks. In this paper, we propose a CWS-specific pre-trained model MetaSeg, which employs a unified architecture and incorporates meta learning algorithm into a multi-criteria pre-training task. Empirical results show that MetaSeg could utilize common prior segmentation knowledge from different existing criteria and alleviate the discrepancy between pre-trained models and downstream CWS tasks. Besides, MetaSeg can achieve new state-of-the-art performance on twelve widely-used CWS datasets and significantly improve model performance in low-resource settings.
Meetings are a key component of human collaboration. As increasing numbers of meetings are recorded and transcribed, meeting summaries have become essential to remind those who may or may not have attended the meetings about the key decisions made and the tasks to be completed. However, it is hard to create a single short summary that covers all the content of a long meeting involving multiple people and topics. In order to satisfy the needs of different types of users, we define a new query-based multi-domain meeting summarization task, where models have to select and summarize relevant spans of meetings in response to a query, and we introduce QMSum, a new benchmark for this task. QMSum consists of 1,808 query-summary pairs over 232 meetings in multiple domains. Besides, we investigate a locate-then-summarize method and evaluate a set of strong summarization baselines on the task. Experimental results and manual analysis reveal that QMSum presents significant challenges in long meeting summarization for future research. Dataset is available at https://github.com/Yale-LILY/QMSum.
Aiming to generate a set of keyphrases, Keyphrase Generation (KG) is a classical task for capturing the central idea from a given document. Based on Seq2Seq models, the previous reinforcement learning framework on KG tasks utilizes the evaluation metrics to further improve the well-trained neural models. However, these KG evaluation metrics such as F1@5 and F1@M are only aware of the exact correctness of predictions on phrase-level and ignore the semantic similarities between similar predictions and targets, which inhibits the model from learning deep linguistic patterns. In response to this problem, we propose a new fine-grained evaluation metric to improve the RL framework, which considers different granularities: token-level F1 score, edit distance, duplication, and prediction quantities. On the whole, the new framework includes two reward functions: the fine-grained evaluation score and the vanilla F1 score. This framework helps the model identifying some partial match phrases which can be further optimized as the exact match ones. Experiments on KG benchmarks show that our proposed training framework outperforms the previous RL training frameworks among all evaluation scores. In addition, our method can effectively ease the synonym problem and generate a higher quality prediction. The source code is available at https://github.com/xuyige/FGRL4KG.
With the continuous upgrading of the summarization systems driven by deep neural networks, researchers have higher requirements on the quality of the generated summaries, which should be not only fluent and informative but also factually correct. As a result, the field of factual evaluation has developed rapidly recently. Despite its initial progress in evaluating generated summaries, the meta-evaluation methodologies of factuality metrics are limited in their opacity, leading to the insufficient understanding of factuality metrics’ relative advantages and their applicability. In this paper, we present an adversarial meta-evaluation methodology that allows us to (i) diagnose the fine-grained strengths and weaknesses of 6 existing top-performing metrics over 24 diagnostic test datasets, (ii) search for directions for further improvement by data augmentation. Our observations from this work motivate us to propose several calls for future research. We make all codes, diagnostic test datasets, trained factuality models available: https://github.com/zide05/AdvFact.
Pre-Trained Models have been widely applied and recently proved vulnerable under backdoor attacks: the released pre-trained weights can be maliciously poisoned with certain triggers. When the triggers are activated, even the fine-tuned model will predict pre-defined labels, causing a security threat. These backdoors generated by the poisoning methods can be erased by changing hyper-parameters during fine-tuning or detected by finding the triggers. In this paper, we propose a stronger weight-poisoning attack method that introduces a layerwise weight poisoning strategy to plant deeper backdoors; we also introduce a combinatorial trigger that cannot be easily detected. The experiments on text classification tasks show that previous defense methods cannot resist our weight-poisoning method, which indicates that our method can be widely applied and may provide hints for future model robustness studies.
Chinese Spelling Check (CSC) is to detect and correct Chinese spelling errors. Many models utilize a predefined confusion set to learn a mapping between correct characters and its visually similar or phonetically similar misuses but the mapping may be out-of-domain. To that end, we propose SpellBERT, a pretrained model with graph-based extra features and independent on confusion set. To explicitly capture the two erroneous patterns, we employ a graph neural network to introduce radical and pinyin information as visual and phonetic features. For better fusing these features with character representations, we devise masked language model alike pre-training tasks. With this feature-rich pre-training, SpellBERT with only half size of BERT can show competitive performance and make a state-of-the-art result on the OCR dataset where most of the errors are not covered by the existing confusion set.
Two important tasks at the intersection of knowledge graphs and natural language processing are graph-to-text (G2T) and text-tograph (T2G) conversion. Due to the difficulty and high cost of data collection, the supervised data available in the two fields are usually on the magnitude of tens of thousands, for example, 18K in the WebNLG 2017 dataset after preprocessing, which is far fewer than the millions of data for other tasks such as machine translation. Consequently, deep learning models for G2T and T2G suffer largely from scarce training data. We present CycleGT, an unsupervised training method that can bootstrap from fully non-parallel graph and text data, and iteratively back translate between the two forms. Experiments on WebNLG datasets show that our unsupervised model trained on the same number of data achieves performance on par with several fully supervised models. Further experiments on the non-parallel GenWiki dataset verify that our method performs the best among unsupervised baselines. This validates our framework as an effective approach to overcome the data scarcity problem in the fields of G2T and T2G.
Text verbalization of knowledge graphs is an important problem with wide application to natural language generation (NLG) systems. It is challenging because the generated text not only needs to be grammatically correct (fluency), but also has to contain the given structured knowledge input (relevance) and meet some other criteria. We develop a plan-and-pretrain approach, 𝒫2, which consists of a relational graph convolutional network (RGCN) planner and the pretrained sequence-tosequence (Seq2Seq) model T5. Specifically, the R-GCN planner first generates an order of the knowledge graph triplets, corresponding to the order that they will be mentioned in text, and then T5 produces the surface realization of the given plan. In the WebNLG+ 2020 Challenge, our submission ranked in 1st place on all automatic and human evaluation criteria of the English RDF-to-text generation task.
Data collection for the knowledge graph-to-text generation is expensive. As a result, research on unsupervised models has emerged as an active field recently. However, most unsupervised models have to use non-parallel versions of existing small supervised datasets, which largely constrain their potential. In this paper, we propose a large-scale, general-domain dataset, GenWiki. Our unsupervised dataset has 1.3M text and graph examples, respectively. With a human-annotated test set, we provide this new benchmark dataset for future research on unsupervised text generation from knowledge graphs.
With the emerging branch of incorporating factual knowledge into pre-trained language models such as BERT, most existing models consider shallow, static, and separately pre-trained entity embeddings, which limits the performance gains of these models. Few works explore the potential of deep contextualized knowledge representation when injecting knowledge. In this paper, we propose the Contextualized Language and Knowledge Embedding (CoLAKE), which jointly learns contextualized representation for both language and knowledge with the extended MLM objective. Instead of injecting only entity embeddings, CoLAKE extracts the knowledge context of an entity from large-scale knowledge bases. To handle the heterogeneity of knowledge context and language context, we integrate them in a unified data structure, word-knowledge graph (WK graph). CoLAKE is pre-trained on large-scale WK graphs with the modified Transformer encoder. We conduct experiments on knowledge-driven tasks, knowledge probing tasks, and language understanding tasks. Experimental results show that CoLAKE outperforms previous counterparts on most of the tasks. Besides, CoLAKE achieves surprisingly high performance on our synthetic task called word-knowledge graph completion, which shows the superiority of simultaneously contextualizing language and knowledge representation.
This paper creates a paradigm shift with regard to the way we build neural extractive summarization systems. Instead of following the commonly used framework of extracting sentences individually and modeling the relationship between sentences, we formulate the extractive summarization task as a semantic text matching problem, in which a source document and candidate summaries will be (extracted from the original text) matched in a semantic space. Notably, this paradigm shift to semantic matching framework is well-grounded in our comprehensive analysis of the inherent gap between sentence-level and summary-level extractors based on the property of the dataset. Besides, even instantiating the framework with a simple form of a matching model, we have driven the state-of-the-art extractive result on CNN/DailyMail to a new level (44.41 in ROUGE-1). Experiments on the other five datasets also show the effectiveness of the matching framework. We believe the power of this matching-based summarization framework has not been fully exploited. To encourage more instantiations in the future, we have released our codes, processed dataset, as well as generated summaries in https://github.com/maszhongming/MatchSum.
As a crucial step in extractive document summarization, learning cross-sentence relations has been explored by a plethora of approaches. An intuitive way is to put them in the graph-based neural network, which has a more complex structure for capturing inter-sentence relationships. In this paper, we present a heterogeneous graph-based neural network for extractive summarization (HETERSUMGRAPH), which contains semantic nodes of different granularity levels apart from sentences. These additional nodes act as the intermediary between sentences and enrich the cross-sentence relations. Besides, our graph structure is flexible in natural extension from a single-document setting to multi-document via introducing document nodes. To our knowledge, we are the first one to introduce different types of nodes into graph-based neural networks for extractive document summarization and perform a comprehensive qualitative analysis to investigate their benefits. The code will be released on Github.
Recently, the character-word lattice structure has been proved to be effective for Chinese named entity recognition (NER) by incorporating the word information. However, since the lattice structure is complex and dynamic, the lattice-based models are hard to fully utilize the parallel computation of GPUs and usually have a low inference speed. In this paper, we propose FLAT: Flat-LAttice Transformer for Chinese NER, which converts the lattice structure into a flat structure consisting of spans. Each span corresponds to a character or latent word and its position in the original lattice. With the power of Transformer and well-designed position encoding, FLAT can fully leverage the lattice information and has an excellent parallel ability. Experiments on four datasets show FLAT outperforms other lexicon-based models in performance and efficiency.
Image captioning is a multimodal problem that has drawn extensive attention in both the natural language processing and computer vision community. In this paper, we present a novel image captioning architecture to better explore semantics available in captions and leverage that to enhance both image representation and caption generation. Our models first construct caption-guided visual relationship graphs that introduce beneficial inductive bias using weakly supervised multi-instance learning. The representation is then enhanced with neighbouring and contextual nodes with their textual and visual features. During generation, the model further incorporates visual relationships using multi-task learning for jointly predicting word and object/predicate tag sequences. We perform extensive experiments on the MSCOCO dataset, showing that the proposed framework significantly outperforms the baselines, resulting in the state-of-the-art performance under a wide range of evaluation metrics. The code of our paper has been made publicly available.
Chinese word segmentation and dependency parsing are two fundamental tasks for Chinese natural language processing. The dependency parsing is defined at the word-level. Therefore word segmentation is the precondition of dependency parsing, which makes dependency parsing suffer from error propagation and unable to directly make use of character-level pre-trained language models (such as BERT). In this paper, we propose a graph-based model to integrate Chinese word segmentation and dependency parsing. Different from previous transition-based joint models, our proposed model is more concise, which results in fewer efforts of feature engineering. Our graph-based joint model achieves better performance than previous joint models and state-of-the-art results in both Chinese word segmentation and dependency parsing. Additionally, when BERT is combined, our model can substantially reduce the performance gap of dependency parsing between joint models and gold-segmented word-based models. Our code is publicly available at https://github.com/fastnlp/JointCwsParser
Multi-criteria Chinese word segmentation (MCCWS) aims to exploit the relations among the multiple heterogeneous segmentation criteria and further improve the performance of each single criterion. Previous work usually regards MCCWS as different tasks, which are learned together under the multi-task learning framework. In this paper, we propose a concise but effective unified model for MCCWS, which is fully-shared for all the criteria. By leveraging the powerful ability of the Transformer encoder, the proposed unified model can segment Chinese text according to a unique criterion-token indicating the output criterion. Besides, the proposed unified model can segment both simplified and traditional Chinese and has an excellent transfer capability. Experiments on eight datasets with different criteria show that our model outperforms our single-criterion baseline model and other multi-criteria models. Source codes of this paper are available on Github.
Neural network-based models augmented with unsupervised pre-trained knowledge have achieved impressive performance on text summarization. However, most existing evaluation methods are limited to an in-domain setting, where summarizers are trained and evaluated on the same dataset. We argue that this approach can narrow our understanding of the generalization ability for different summarization systems. In this paper, we perform an in-depth analysis of characteristics of different datasets and investigate the performance of different summarization models under a cross-dataset setting, in which a summarizer trained on one corpus will be evaluated on a range of out-of-domain corpora. A comprehensive study of 11 representative summarization systems on 5 datasets from different domains reveals the effect of model architectures and generation ways (i.e. abstractive and extractive) on model generalization ability. Further, experimental results shed light on the limitations of existing summarizers. Brief introduction and supplementary code can be found in https://github.com/zide05/CDEvalSumm.
Reverse dictionary is the task to find the proper target word given the word description. In this paper, we tried to incorporate BERT into this task. However, since BERT is based on the byte-pair-encoding (BPE) subword encoding, it is nontrivial to make BERT generate a word given the description. We propose a simple but effective method to make BERT generate the target word for this specific task. Besides, the cross-lingual reverse dictionary is the task to find the proper target word described in another language. Previous models have to keep two different word embeddings and learn to align these embeddings. Nevertheless, by using the Multilingual BERT (mBERT), we can efficiently conduct the cross-lingual reverse dictionary with one subword embedding, and the alignment between languages is not necessary. More importantly, mBERT can achieve remarkable cross-lingual reverse dictionary performance even without the parallel corpus, which means it can conduct the cross-lingual reverse dictionary with only corresponding monolingual data. Code is publicly available at https://github.com/yhcc/BertForRD.git.
We investigate the following question for machine translation (MT): can we develop a single universal MT model to serve as the common seed and obtain derivative and improved models on arbitrary language pairs? We propose mRASP, an approach to pre-train a universal multilingual neural machine translation model. Our key idea in mRASP is its novel technique of random aligned substitution, which brings words and phrases with similar meanings across multiple languages closer in the representation space. We pre-train a mRASP model on 32 language pairs jointly with only public datasets. The model is then fine-tuned on downstream language pairs to obtain specialized MT models. We carry out extensive experiments on 42 translation directions across a diverse settings, including low, medium, rich resource, and as well as transferring to exotic language pairs. Experimental results demonstrate that mRASP achieves significant performance improvement compared to directly training on those target pairs. It is the first time to verify that multiple lowresource language pairs can be utilized to improve rich resource MT. Surprisingly, mRASP is even able to improve the translation quality on exotic languages that never occur in the pretraining corpus. Code, data, and pre-trained models are available at https://github.com/linzehui/mRASP.
Adversarial attacks for discrete data (such as texts) have been proved significantly more challenging than continuous data (such as images) since it is difficult to generate adversarial samples with gradient-based methods. Current successful attack methods for texts usually adopt heuristic replacement strategies on the character or word level, which remains challenging to find the optimal solution in the massive space of possible combinations of replacements while preserving semantic consistency and language fluency. In this paper, we propose BERT-Attack, a high-quality and effective method to generate adversarial samples using pre-trained masked language models exemplified by BERT. We turn BERT against its fine-tuned models and other deep neural models in downstream tasks so that we can successfully mislead the target models to predict incorrectly. Our method outperforms state-of-the-art attack strategies in both success rate and perturb percentage, while the generated adversarial samples are fluent and semantically preserved. Also, the cost of calculation is low, thus possible for large-scale generations. The code is available at https://github.com/LinyangLee/BERT-Attack.
The recent years have seen remarkable success in the use of deep neural networks on text summarization. However, there is no clear understanding of why they perform so well, or how they might be improved. In this paper, we seek to better understand how neural extractive summarization systems could benefit from different types of model architectures, transferable knowledge and learning schemas. Besides, we find an effective way to improve the current framework and achieve the state-of-the-art result on CNN/DailyMail by a large margin based on our observations and analysis. Hopefully, our work could provide more hints for future research on extractive summarization.
Disentangling the content and style in the latent space is prevalent in unpaired text style transfer. However, two major issues exist in most of the current neural models. 1) It is difficult to completely strip the style information from the semantics for a sentence. 2) The recurrent neural network (RNN) based encoder and decoder, mediated by the latent representation, cannot well deal with the issue of the long-term dependency, resulting in poor preservation of non-stylistic semantic content. In this paper, we propose the Style Transformer, which makes no assumption about the latent representation of source sentence and equips the power of attention mechanism in Transformer to achieve better style transfer and better content preservation.
Aspect-based sentiment analysis (ABSA), which aims to identify fine-grained opinion polarity towards a specific aspect, is a challenging subtask of sentiment analysis (SA). In this paper, we construct an auxiliary sentence from the aspect and convert ABSA to a sentence-pair classification task, such as question answering (QA) and natural language inference (NLI). We fine-tune the pre-trained model from BERT and achieve new state-of-the-art results on SentiHood and SemEval-2014 Task 4 datasets. The source codes are available at https://github.com/HSLCY/ABSA-BERT-pair.
Although Transformer has achieved great successes on many NLP tasks, its heavy structure with fully-connected attention connections leads to dependencies on large training data. In this paper, we present Star-Transformer, a lightweight alternative by careful sparsification. To reduce model complexity, we replace the fully-connected structure with a star-shaped topology, in which every two non-adjacent nodes are connected through a shared relay node. Thus, complexity is reduced from quadratic to linear, while preserving the capacity to capture both local composition and long-range dependency. The experiments on four tasks (22 datasets) show that Star-Transformer achieved significant improvements against the standard Transformer for the modestly sized datasets.
Chinese is a logographic writing system, and the shape of Chinese characters contain rich syntactic and semantic information. In this paper, we propose a model to learn Chinese word embeddings via three-level composition: (1) a convolutional neural network to extract the intra-character compositionality from the visual shape of a character; (2) a recurrent neural network with self-attention to compose character representation into word embeddings; (3) the Skip-Gram framework to capture non-compositionality directly from the contextual information. Evaluations demonstrate the superior performance of our model on four tasks: word similarity, sentiment analysis, named entity recognition and part-of-speech tagging.
Word Sense Disambiguation (WSD) aims to find the exact sense of an ambiguous word in a particular context. Traditional supervised methods rarely take into consideration the lexical resources like WordNet, which are widely utilized in knowledge-based methods. Recent studies have shown the effectiveness of incorporating gloss (sense definition) into neural networks for WSD. However, compared with traditional word expert supervised methods, they have not achieved much improvement. In this paper, we focus on how to better leverage gloss knowledge in a supervised neural WSD system. We construct context-gloss pairs and propose three BERT based models for WSD. We fine-tune the pre-trained BERT model and achieve new state-of-the-art results on WSD task.
In this paper, we take stock of the current state of summarization datasets and explore how different factors of datasets influence the generalization behaviour of neural extractive summarization models. Specifically, we first propose several properties of datasets, which matter for the generalization of summarization models. Then we build the connection between priors residing in datasets and model designs, analyzing how different properties of datasets influence the choices of model structure design and training methods. Finally, by taking a typical dataset as an example, we rethink the process of the model design based on the experience of the above analysis. We demonstrate that when we have a deep understanding of the characteristics of datasets, a simple approach can bring significant improvements to the existing state-of-the-art model.
This paper describes Fudan’s submission to CoNLL 2018’s shared task Universal Dependency Parsing. We jointly train models when two languages are similar according to linguistic typology and then ensemble the models using a simple re-parse algorithm. We outperform the baseline method by 4.4% (2.1%) on average on development (test) set in CoNLL 2018 UD Shared Task.
While much progress has been made in how to encode a text sequence into a sequence of vectors, less attention has been paid to how to aggregate these preceding vectors (outputs of RNN/CNN) into fixed-size encoding vector. Usually, a simple max or average pooling is used, which is a bottom-up and passive way of aggregation and lack of guidance by task information. In this paper, we propose an aggregation mechanism to obtain a fixed-size encoding with a dynamic routing policy. The dynamic routing policy is dynamically deciding that what and how much information need be transferred from each word to the final encoding of the text sequence. Following the work of Capsule Network, we design two dynamic routing policies to aggregate the outputs of RNN/CNN encoding layer into a final encoding vector. Compared to the other aggregation methods, dynamic routing can refine the messages according to the state of final encoding vector. Experimental results on five text classification tasks show that our method outperforms other aggregating models by a significant margin. Related source code is released on our github page. Related source code is released on our github page.
Attention-based neural models have achieved great success in natural language inference (NLI). In this paper, we propose the Convolutional Interaction Network (CIN), a general model to capture the interaction between two sentences, which can be an alternative to the attention mechanism for NLI. Specifically, CIN encodes one sentence with the filters dynamically generated based on another sentence. Since the filters may be designed to have various numbers and sizes, CIN can capture more complicated interaction patterns. Experiments on three large datasets demonstrate CIN’s efficacy.
Idioms are peculiar linguistic constructions that impose great challenges for representing the semantics of language, especially in current prevailing end-to-end neural models, which assume that the semantics of a phrase or sentence can be literally composed from its constitutive words. In this paper, we propose an idiom-aware distributed semantic model to build representation of sentences on the basis of understanding their contained idioms. Our models are grounded in the literal-first psycholinguistic hypothesis, which can adaptively learn semantic compositionality of a phrase literally or idiomatically. To better evaluate our models, we also construct an idiom-enriched sentiment classification dataset with considerable scale and abundant peculiarities of idioms. The qualitative and quantitative experimental analyses demonstrate the efficacy of our models.
Neural network models have shown their promising opportunities for multi-task learning, which focus on learning the shared layers to extract the common and task-invariant features. However, in most existing approaches, the extracted shared features are prone to be contaminated by task-specific features or the noise brought by other tasks. In this paper, we propose an adversarial multi-task learning framework, alleviating the shared and private latent feature spaces from interfering with each other. We conduct extensive experiments on 16 different text classification tasks, which demonstrates the benefits of our approach. Besides, we show that the shared knowledge learned by our proposed model can be regarded as off-the-shelf knowledge and easily transferred to new tasks. The datasets of all 16 tasks are publicly available at http://nlp.fudan.edu.cn/data/.
Different linguistic perspectives causes many diverse segmentation criteria for Chinese word segmentation (CWS). Most existing methods focus on improve the performance for each single criterion. However, it is interesting to exploit these different criteria and mining their common underlying knowledge. In this paper, we propose adversarial multi-criteria learning for CWS by integrating shared knowledge from multiple heterogeneous segmentation criteria. Experiments on eight corpora with heterogeneous segmentation criteria show that the performance of each corpus obtains a significant improvement, compared to single-criterion learning. Source codes of this paper are available on Github.