While long-context large language models (LLMs) exhibit remarkable document processing capabilities, their prohibitively high training costs often hinder customized applications. To mitigate this issue, we propose __Sequential Chunk-wise Optimization (SeCO)__, a memory-efficient training paradigm that partitions lengthy inputs into manageable chunks. Each chunk independently constructs its computational graph and performs localized backpropagation, ensuring that only one chunk’s forward activations are stored in memory. Building on SeCO, we further introduce __Sparse Chunk-wise Optimization (SpaCO)__, which reduces computational overhead by selectively propagating gradients to specific chunks and incorporates a carefully designed compensation factor to ensure unbiased gradient estimation. SpaCO decouples the computational cost of backpropagation from the context length, enabling training time to gradually converge to inference time as sequences become longer. Implemented as lightweight training wrappers, both SeCO and SpaCO offer substantial practical benefits. For example, when fine-tuning an 8B model with LoRA on a single RTX 3090 GPU, SeCO expands maximum sequence length from 1K to 16K tokens, while SpaCO demonstrates accelerated training speed—achieving up to 3× faster than SeCO under the same experimental setup. These innovations provide new insights into optimizing long-context models, making them more accessible for practical applications. We have open-sourced the code at https://anonymous.4open.science/r/seco-CCBD.

Retrieval-augmented generation (RAG) enhances large language models (LLMs) by integrating external knowledge. A critical yet underexplored challenge in RAG is document segmentation, also known as document chunking. Existing widely-used rule-based chunking methods usually lead to suboptimal splits, where overly large chunks introduce irrelevant information and small chunks lack semantic coherence. Existing semantic-based approaches either require costly LLM calls or fail to adaptively group contextually related sentences. To address these limitations, we propose PIC, Pseudo-Instruction for document Chunking), a simple yet effective method that leverages document summaries as pseudo-instructions to guide chunking. By computing semantic similarity between sentences and the summary, PIC dynamically groups sentences into chunks that align with the document’s key themes, ensuring semantic completeness and relevance to potential user instructions. Experiments on multiple open-domain question-answering benchmarks demonstrate that PIC can significantly improve retrieval accuracy (Hits@k) and end-to-end QA performance (Exact Match) without any additional training.

Knowledge Graph Completion (KGC), which aims to infer missing or incomplete facts, is a crucial task for KGs. However, integrating the vital structural information of KGs into Large Language Models (LLMs) and outputting predictions deterministically remains challenging. To address this, we propose a new method called GLTW, which encodes the structural information of KGs and merges it with LLMs to enhance KGC performance. Specifically, we introduce an improved Graph Transformer (iGT) that effectively encodes subgraphs with both local and global structural information and inherits the characteristics of language model, bypassing training from scratch. Also, we develop a subgraph-based multi-classification training objective, using all entities within KG as classification objects, to boost learning efficiency. Importantly, we combine iGT with an LLM that takes KG language prompts as input. Our extensive experiments on various KG datasets show that GLTW achieves significant performance gains compared to SOTA baselines.

Vision-language (VL) understanding tasks evaluate models’ comprehension of complex visual scenes through multiple-choice questions. However, we have identified two dataset biases that models can exploit as shortcuts to resolve various VL tasks correctly without proper understanding. The first type of dataset bias is Unbalanced Matching bias, where the correct answer overlaps the question and image more than the incorrect answers. The second type of dataset bias is Distractor Similarity bias, where incorrect answers are overly dissimilar to the correct answer but significantly similar to other incorrect answers within the same sample. To address these dataset biases, we first propose Adversarial Data Synthesis (ADS) to generate synthetic training and debiased evaluation data. We then introduce Intra-sample Counterfactual Training (ICT) to assist models in utilizing the synthesized training data, particularly the counterfactual data, via focusing on intra-sample differentiation. Extensive experiments demonstrate the effectiveness of ADS and ICT in consistently improving model performance across different benchmarks, even in domain-shifted scenarios.

Visual commonsense understanding requires Vision Language (VL) models to not only understand image and text but also cross-reference in-between to fully integrate and achieve comprehension of the visual scene described. Recently, various approaches have been developed and have achieved high performance on visual commonsense benchmarks. However, it is unclear whether the models really understand the visual scene and underlying commonsense knowledge due to limited evaluation data resources. To provide an in-depth analysis, we present a Multimodal Evaluation (ME) pipeline to automatically generate question-answer pairs to test models’ understanding of the visual scene, text, and related knowledge. We then take a step further to show that training with the ME data boosts the model’s performance in standard VCR evaluation. Lastly, our in-depth analysis and comparison reveal interesting findings: (1) semantically low-level information can assist the learning of high-level information but not the opposite; (2) visual information is generally under utilization compared with text.

Recently, powerful Transformer architectures have proven superior in generating high-quality sentences. Nevertheless, these models tend to produce dull high-frequency phrases, severely hurting the diversity and novelty of generated text. In this work, we dig into the intrinsic mechanism of this problem and found that sparser attention values in Transformer could improve diversity. To understand such a phenomenon, we first conduct both empirical and theoretical analysis and then attribute it to representation degeneration caused by the attentive mixture of the hidden states during training. We term this process the Trap of Mediocrity. To escape from such a trap, we introduce a novel attention regularization loss to control the sharpness of the attention distribution, which is transparent to model structures and can be easily implemented within 20 lines of python code. We prove that this method could be mathematically regarded as learning a Bayesian approximation of posterior attention. Experiments show that our method improved the diversity and novelty of the generated text while maintaining comparable quality on a variety of conditional and unconditional generation tasks.

Variational Auto-Encoder (VAE) has been widely adopted in text generation. Among many variants, recurrent VAE learns token-wise latent variables with each conditioned on the preceding ones, which captures sequential variability better in the era of RNN. However, it is unclear how to incorporate such recurrent dynamics into the recently dominant Transformer due to its parallelism. In this work, we propose TRACE, a Transformer-based recurrent VAE structure. TRACE imposes recurrence on segment-wise latent variables with arbitrarily separated text segments and constructs the posterior distribution with residual parameterization. Besides, we design an acceleration method by approximating idempotent matrices, which allows parallelism while maintaining the conditional dependence of latent variables. We demonstrate that TRACE could deduce a non-zero lower bound of the KL term and enhance the entanglement of each segment and preceding latent variables, providing a theoretical guarantee of generation diversity. Experiments on two unconditional and one conditional generation task show that TRACE achieves significantly improved diversity while maintaining satisfactory generation quality.

The past several years have witnessed Variational Auto-Encoder’s superiority in various text generation tasks. However, due to the sequential nature of the text, auto-regressive decoders tend to ignore latent variables and then reduce to simple language models, known as the KL vanishing problem, which would further deteriorate when VAE is combined with Transformer-based structures. To ameliorate this problem, we propose Della, a novel variational Transformer framework. Della learns a series of layer-wise latent variables with each inferred from those of lower layers and tightly coupled with the hidden states by low-rank tensor product. In this way, Della forces these posterior latent variables to be fused deeply with the whole computation path and hence incorporate more information. We theoretically demonstrate that our method can be regarded as entangling latent variables to avoid posterior information decrease through layers, enabling Della to get higher non-zero KL values even without any annealing or thresholding tricks. Experiments on four unconditional and three conditional generation tasks show that Della could better alleviate KL vanishing and improve both quality and diversity compared to several strong baselines.

中国古代诗歌是人类文化的瑰宝,其短小精悍的语言却能表达出极其丰富的含义和主题,从古至今吸引了无数的爱好者的欣赏。本文以超过锸锰万首古诗为研究对象,基于BPE算法,按照共现频率对古诗集进行分词,以便于下游任务对古诗的语义进行更准确的理解,我们还将分词后的古诗语料利用隐狄利克雷分配(LDA)模型进行了主题分析。通过比较、调整主题的数量得到了准确度较高的主题模型。更进一步,我们还对语料中的绝句和律诗逐句套用了主题模型,得到了一首诗内部的主题转移矩阵,并进行了一些相关的分析。最后,我们利用了简单的控制码方法将主题模型嵌入到诗歌生成模型中,实现了主题可控的诗歌生成,同时检验了我们训练的主题模型的有效性。

Research on the automatic generation of poetry, the treasure of human culture, has lasted for decades. Most existing systems, however, are merely model-oriented, which input some user-specified keywords and directly complete the generation process in one pass, with little user participation. We believe that the machine, being a collaborator or an assistant, should not replace human beings in poetic creation. Therefore, we proposed Jiuge, a human-machine collaborative Chinese classical poetry generation system. Unlike previous systems, Jiuge allows users to revise the unsatisfied parts of a generated poem draft repeatedly. According to the revision, the poem will be dynamically updated and regenerated. After the revision and modification procedure, the user can write a satisfying poem together with Jiuge system collaboratively. Besides, Jiuge can accept multi-modal inputs, such as keywords, plain text or images. By exposing the options of poetry genres, styles and revision modes, Jiuge, acting as a professional assistant, allows constant and active participation of users in poetic creation.

Poetry is one of the most beautiful forms of human language art. As a crucial step towards computer creativity, automatic poetry generation has drawn researchers’ attention for decades. In recent years, some neural models have made remarkable progress in this task. However, they are all based on maximum likelihood estimation, which only learns common patterns of the corpus and results in loss-evaluation mismatch. Human experts evaluate poetry in terms of some specific criteria, instead of word-level likelihood. To handle this problem, we directly model the criteria and use them as explicit rewards to guide gradient update by reinforcement learning, so as to motivate the model to pursue higher scores. Besides, inspired by writing theories, we propose a novel mutual reinforcement learning schema. We simultaneously train two learners (generators) which learn not only from the teacher (rewarder) but also from each other to further improve performance. We experiment on Chinese poetry. Based on a strong basic model, our method achieves better results and outperforms the current state-of-the-art method.

The ability to write diverse poems in different styles under the same poetic imagery is an important characteristic of human poetry writing. Most previous works on automatic Chinese poetry generation focused on improving the coherency among lines. Some work explored style transfer but suffered from expensive expert labeling of poem styles. In this paper, we target on stylistic poetry generation in a fully unsupervised manner for the first time. We propose a novel model which requires no supervised style labeling by incorporating mutual information, a concept in information theory, into modeling. Experimental results show that our model is able to generate stylistic poems without losing fluency and coherency.