Previous work has showcased the intriguing capability of large language models (LLMs) in retrieving facts and processing context knowledge. However, only limited research exists on the layer-wise capability of LLMs to encode knowledge, which challenges our understanding of their internal mechanisms. In this paper, we devote the first attempt to investigate the layer-wise capability of LLMs through probing tasks. We leverage the powerful generative capability of ChatGPT to construct probing datasets, providing diverse and coherent evidence corresponding to various facts. We employ \mathcal V-usable information as the validation metric to better reflect the capability in encoding context knowledge across different layers. Our experiments on conflicting and newly acquired knowledge show that LLMs: (1) prefer to encode more context knowledge in the upper layers; (2) primarily encode context knowledge within knowledge-related entity tokens at lower layers while progressively expanding more knowledge within other tokens at upper layers; and (3) gradually forget the earlier context knowledge retained within the intermediate layers when provided with irrelevant evidence. Code is publicly available at https://github.com/Jometeorie/probing_llama.

Large language models (LLMs) have shown tremendous success in following user instructions and generating helpful responses. Nevertheless, their robustness is still far from optimal, as they may generate significantly inconsistent responses due to minor changes in the verbalized instructions. Recent literature has explored this inconsistency issue, highlighting the importance of continued improvement in the robustness of response generation. However, systematic analysis and solutions are still lacking. In this paper, we quantitatively define the inconsistency problem and propose a two-stage training framework consisting of instruction-augmented supervised fine-tuning and consistency alignment training. The first stage helps a model generalize on following instructions via similar instruction augmentations. In the second stage, we improve the diversity and help the model understand which responses are more aligned with human expectations by differentiating subtle differences in similar responses. The training process is accomplished by self-rewards inferred from the trained model at the first stage without referring to external human preference resources. We conduct extensive experiments on recent publicly available LLMs on instruction-following tasks and demonstrate the effectiveness of our training framework.

Large Language Models (LLMs) have shown great potential in Natural Language Processing (NLP) tasks.However, recent literature reveals that LLMs hallucinate intermittently, which impedes their reliability for further utilization. In this paper, we propose a novel self-detection method to detect which questions an LLM does not know.Our proposal is empirical and applicable for continually upgrading LLMs compared with state-of-the-art methods. Specifically, we examine the divergence of the LLM’s behaviors on different verbalizations for a question and examine the atypicality of the verbalized input. We combine the two components to identify whether the model generates a non-factual response to the question. The above components can be accomplished by utilizing the LLM itself without referring to any other external resources. We conduct comprehensive experiments and demonstrate the effectiveness of our method for recently released LLMs involving Llama 2, Vicuna, ChatGPT, and GPT-4 across factoid question-answering, arithmetic reasoning, and commonsense reasoning tasks.

This paper revisits tokenization from a theoretical perspective, and argues for the necessity of a constructivist approach to tokenization for semantic parsing and modeling language acquisition. We consider two problems: (1) (semi-) automatically converting existing lexicalist annotations, e.g. those of the Penn TreeBank, into constructivist annotations, and (2) automatic tokenization of raw texts. We demonstrate that (1) a heuristic rule-based constructivist tokenizer is able to yield relatively satisfactory accuracy when gold standard Penn TreeBank part-of-speech tags are available, but that some manual annotations are still necessary to obtain gold standard results, and (2) a neural tokenizer is able to provide accurate automatic constructivist tokenization results from raw character sequences. Our research output also includes a set of high-quality morpheme-tokenized corpora, which enable the training of computational models that more closely align with language comprehension and acquisition.

Knowledge selection is the key in knowledge-grounded dialogues (KGD), which aims to select an appropriate knowledge snippet to be used in the utterance based on dialogue history. Previous studies mainly employ the classification approach to classify each candidate snippet as “relevant” or “irrelevant” independently. However, such approaches neglect the interactions between snippets, leading to difficulties in inferring the meaning of snippets. Moreover, they lack modeling of the discourse structure of dialogue-knowledge interactions. We propose a simple yet effective generative approach for knowledge selection, called GenKS. GenKS learns to select snippets by generating their identifiers with a sequence-to-sequence model. GenKS therefore captures intra-knowledge interaction inherently through attention mechanisms. Meanwhile, we devise a hyperlink mechanism to model the dialogue-knowledge interactions explicitly. We conduct experiments on three benchmark datasets, and verify GenKS achieves the best results on both knowledge selection and response generation.

Dialogue assessment plays a critical role in the development of open-domain dialogue systems. Existing work are uncapable of providing an end-to-end and human-epistemic assessment dataset, while they only provide sub-metrics like coherence or the dialogues are conversed between annotators far from real user settings. In this paper, we release a large-scale dialogue quality assessment dataset (DiQAD), for automatically assessing open-domain dialogue quality. Specifically, we (1) establish the assessment criteria based on the dimensions conforming to human judgements on dialogue qualities, and (2) annotate large-scale dialogues that conversed between real users based on these annotation criteria, which contains around 100,000 dialogues. We conduct several experiments and report the performances of the baselines as the benchmark on DiQAD. The dataset is openly accessible at https://github.com/yukunZhao/Dataset_Dialogue_quality_evaluation.

Large Language Models (LLMs) have demonstrated remarkable zero-shot generalization across various language-related tasks, including search engines. However, existing work utilizes the generative ability of LLMs for Information Retrieval (IR) rather than direct passage ranking. The discrepancy between the pre-training objectives of LLMs and the ranking objective poses another challenge. In this paper, we first investigate generative LLMs such as ChatGPT and GPT-4 for relevance ranking in IR. Surprisingly, our experiments reveal that properly instructed LLMs can deliver competitive, even superior results to state-of-the-art supervised methods on popular IR benchmarks. Furthermore, to address concerns about data contamination of LLMs, we collect a new test set called NovelEval, based on the latest knowledge and aiming to verify the model’s ability to rank unknown knowledge. Finally, to improve efficiency in real-world applications, we delve into the potential for distilling the ranking capabilities of ChatGPT into small specialized models using a permutation distillation scheme. Our evaluation results turn out that a distilled 440M model outperforms a 3B supervised model on the BEIR benchmark. The code to reproduce our results is available at www.github.com/sunnweiwei/RankGPT.

In open-domain question answering, due to the ambiguity of questions, multiple plausible answers may exist. To provide feasible answers to an ambiguous question,one approach is to directly predict all valid answers, but this can struggle with balancing relevance and diversity. An alternative is to gather candidate answers and aggregate them, but this method can be computationally costly and may neglect dependencies among answers. In this paper, we present AmbigPrompt to address the imperfections of existing approaches to answering ambiguous questions. Specifically, we integrate an answering model with a prompting model in an iterative manner. The prompting model adaptively tracks the reading process and progressively triggers the answering model to compose distinct and relevant answers. Additionally, we develop a task-specific post-pretraining approach for both the answering model and the prompting model, which greatly improves the performance of our framework. Empirical studies on two commonly-used open benchmarks show that AmbigPrompt achieves state-of-the-art or competitive results while using less memory and having a lower inference latency than competing approaches. Additionally, AmbigPrompt also performs well in low-resource settings.

Evaluating open-domain dialogue systems is challenging for reasons such as the one-to-many problem, i.e., many appropriate responses other than just the golden response. As of now, automatic evaluation methods need better consistency with humans, while reliable human evaluation can be time- and cost-intensive. To this end, we propose the Reference-Assisted Dialogue Evaluation (RADE) approach under the multi-task learning framework, which leverages the pre-created utterance as reference other than the gold response to relief the one-to-many problem. Specifically, RADE explicitly compares reference and the candidate response to predict their overall scores. Moreover, an auxiliary response generation task enhances prediction via a shared encoder. To support RADE, we extend three datasets with additional rated responses other than just a golden response by human annotation. Experiments on our three datasets and two existing benchmarks demonstrate the effectiveness of our method, where Pearson, Spearman, and Kendall correlations with human evaluation outperform state-of-the-art baselines.

Universal Semantic Tagging aims to provide lightweight unified analysis for all languages at the word level. Though the proposed annotation scheme is conceptually promising, the feasibility is only examined in four Indo–European languages. This paper is concerned with extending the annotation scheme to handle Mandarin Chinese and empirically study the plausibility of unifying meaning representations for multiple languages. We discuss a set of language-specific semantic phenomena, propose new annotation specifications and build a richly annotated corpus. The corpus consists of 1100 English–Chinese parallel sentences, where compositional semantic analysis is available for English, and another 1000 Chinese sentences which has enriched syntactic analysis. By means of the new annotations, we also evaluate a series of neural tagging models to gauge how successful semantic tagging can be: accuracies of 92.7% and 94.6% are obtained for Chinese and English respectively. The English tagging performance is remarkably better than the state-of-the-art by 7.7%.

This paper studies Negation Scope Resolution (NSR) for Chinese as a Second Language (CSL), which shows many unique characteristics that distinguish itself from “standard” Chinese. We annotate a new moderate-sized corpus that covers two background L1 languages, viz. English and Japanese. We build a neural NSR system, which achieves a new state-of-the-art accuracy on English benchmark data. We leverage this system to gauge how successful NSR for CSL can be. Different native language backgrounds of language learners result in unequal cross-lingual transfer, which has a significant impact on processing second language data. In particular, manual annotation, empirical evaluation and error analysis indicate two non-obvious facts: 1) L2-Chinese, L1-Japanese data are more difficult to analyze and thus annotate than L2-Chinese, L1-English data; 2) computational models trained on L2-Chinese, L1-Japanese data perform better than models trained on L2-Chinese, L1-English data.

In this work, we present a phenomenon-oriented comparative analysis of the two dominant approaches in English Resource Semantic (ERS) parsing: classic, knowledge-intensive and neural, data-intensive models. To reflect state-of-the-art neural NLP technologies, a factorization-based parser is introduced that can produce Elementary Dependency Structures much more accurately than previous data-driven parsers. We conduct a suite of tests for different linguistic phenomena to analyze the grammatical competence of different parsers, where we show that, despite comparable performance overall, knowledge- and data-intensive models produce different types of errors, in a way that can be explained by their theoretical properties. This analysis is beneficial to in-depth evaluation of several representative parsing techniques and leads to new directions for parser development.

A key problem in processing graph-based meaning representations is graph parsing, i.e. computing all possible derivations of a given graph according to a (competence) grammar. We demonstrate, for the first time, that exact graph parsing can be efficient for large graphs and with large Hyperedge Replacement Grammars (HRGs). The advance is achieved by exploiting locality as terminal edge-adjacency in HRG rules. In particular, we highlight the importance of 1) a terminal edge-first parsing strategy, 2) a categorization of a subclass of HRG, i.e. what we call Weakly Regular Graph Grammar, and 3) distributing argument-structures to both lexical and phrasal rules.

We propose variable-in-situ logico-semantic graphs to bridge the gap between semantic graph and logical form parsing. The new type of graph-based meaning representation allows us to include analysis for scope-related phenomena, such as quantification, negation and modality, in a way that is consistent with the state-of-the-art underspecification approach. Moreover, the well-formedness of such a graph is clear, since model-theoretic interpretation is available. We demonstrate the effectiveness of this new perspective by developing a new state-of-the-art semantic parser for English Resource Semantics. At the core of this parser is a novel neural graph rewriting system which combines the strengths of Hyperedge Replacement Grammar, a knowledge-intensive model, and Graph Neural Networks, a data-intensive model. Our parser achieves an accuracy of 92.39% in terms of elementary dependency match, which is a 2.88 point improvement over the best data-driven model in the literature. The output of our parser is highly coherent: at least 91% graphs are valid, in that they allow at least one sound scope-resolved logical form.

This paper is concerned with semantic parsing for English as a second language (ESL). Motivated by the theoretical emphasis on the learning challenges that occur at the syntax-semantics interface during second language acquisition, we formulate the task based on the divergence between literal and intended meanings. We combine the complementary strengths of English Resource Grammar, a linguistically-precise hand-crafted deep grammar, and TLE, an existing manually annotated ESL UD-TreeBank with a novel reranking model. Experiments demonstrate that in comparison to human annotations, our method can obtain a very promising SemBanking quality. By means of the newly created corpus, we evaluate state-of-the-art semantic parsing as well as grammatical error correction models. The evaluation profiles the performance of neural NLP techniques for handling ESL data and suggests some research directions.

Natural Language Processing (NLP) tasks are usually performed word by word on textual inputs. We can use arbitrary symbols to represent the linguistic meaning of a word and use these symbols as inputs. As “alternatives” to a text representation, we introduce Soundex, MetaPhone, NYSIIS, logogram to NLP, and develop fixed-output-length coding and its extension using Huffman coding. Each of those codings combines different character/digital sequences and constructs a new vocabulary based on codewords. We find that the integration of those codewords with text provides more reliable inputs to Neural-Network-based NLP systems through redundancy than text-alone inputs. Experiments demonstrate that our approach outperforms the state-of-the-art models on the application of machine translation, language modeling, and part-of-speech tagging. The source code is available at https://github.com/abdulrafae/coding_nmt.

This paper describes the systems of the CUNY-PKU team in SemEval 2019 Task 1: Cross-lingual Semantic Parsing with UCCA. We introduce a novel model by applying a cascaded MLP and BiLSTM model. Then, we ensemble multiple system-outputs by reparsing. In particular, we introduce a new decoding algorithm for building the UCCA representation. Our system won the first place in one track (French-20K-Open), second places in four tracks (English-Wiki-Open, English-20K-Open, German-20K-Open, and German-20K-Closed), and third place in one track (English-20K-Closed), among all seven tracks.

This tutorial is on representing and processing sentence meaning in the form of labeled directed graphs. The tutorial will (a) briefly review relevant background in formal and linguistic semantics; (b) semi-formally define a unified abstract view on different flavors of semantic graphs and associated terminology; (c) survey common frameworks for graph-based meaning representation and available graph banks; and (d) offer a technical overview of a representative selection of different parsing approaches.

We report our work on building linguistic resources and data-driven parsers in the grammatical relation (GR) analysis for Mandarin Chinese. Chinese, as an analytic language, encodes grammatical information in a highly configurational rather than morphological way. Accordingly, it is possible and reasonable to represent almost all grammatical relations as bilexical dependencies. In this work, we propose to represent grammatical information using general directed dependency graphs. Both only-local and rich long-distance dependencies are explicitly represented. To create high-quality annotations, we take advantage of an existing TreeBank, namely, Chinese TreeBank (CTB), which is grounded on the Government and Binding theory. We define a set of linguistic rules to explore CTB’s implicit phrase structural information and build deep dependency graphs. The reliability of this linguistically motivated GR extraction procedure is highlighted by manual evaluation. Based on the converted corpus, data-driven, including graph- and transition-based, models are explored for Chinese GR parsing. For graph-based parsing, a new perspective, graph merging, is proposed for building flexible dependency graphs: constructing complex graphs via constructing simple subgraphs. Two key problems are discussed in this perspective: (1) how to decompose a complex graph into simple subgraphs, and (2) how to combine subgraphs into a coherent complex graph. For transition-based parsing, we introduce a neural parser based on a list-based transition system. We also discuss several other key problems, including dynamic oracle and beam search for neural transition-based parsing. Evaluation gauges how successful GR parsing for Chinese can be by applying data-driven models. The empirical analysis suggests several directions for future study.

We design, implement and evaluate two semantic parsers, which represent factorization- and composition-based approaches respectively, for Elementary Dependency Structures (EDS) at the CoNLL 2019 Shared Task on Cross-Framework Meaning Representation Parsing. The detailed evaluation of the two parsers gives us a new perception about parsing into linguistically enriched meaning representations: current neural EDS parsers are able to reach an accuracy at the inter-annotator agreement level in the same-epoch-and-domain setup.

We present experiments for cross-domain semantic dependency analysis with a neural Maximum Subgraph parser. Our parser targets 1-endpoint-crossing, pagenumber-2 graphs which are a good fit to semantic dependency graphs, and utilizes an efficient dynamic programming algorithm for decoding. For disambiguation, the parser associates words with BiLSTM vectors and utilizes these vectors to assign scores to candidate dependencies. We conduct experiments on the data sets from SemEval 2015 as well as Chinese CCGBank. Our parser achieves very competitive results for both English and Chinese. To improve the parsing performance on cross-domain texts, we propose a data-oriented method to explore the linguistic generality encoded in English Resource Grammar, which is a precisionoriented, hand-crafted HPSG grammar, in an implicit way. Experiments demonstrate the effectiveness of our data-oriented method across a wide range of conditions.

This paper studies semantic parsing for interlanguage (L2), taking semantic role labeling (SRL) as a case task and learner Chinese as a case language. We first manually annotate the semantic roles for a set of learner texts to derive a gold standard for automatic SRL. Based on the new data, we then evaluate three off-the-shelf SRL systems, i.e., the PCFGLA-parser-based, neural-parser-based and neural-syntax-agnostic systems, to gauge how successful SRL for learner Chinese can be. We find two non-obvious facts: 1) the L1-sentence-trained systems performs rather badly on the L2 data; 2) the performance drop from the L1 data to the L2 data of the two parser-based systems is much smaller, indicating the importance of syntactic parsing in SRL for interlanguages. Finally, the paper introduces a new agreement-based model to explore the semantic coherency information in the large-scale L2-L1 parallel data. We then show such information is very effective to enhance SRL for learner texts. Our model achieves an F-score of 72.06, which is a 2.02 point improvement over the best baseline.

We demonstrate that an SHRG-based parser can produce semantic graphs much more accurately than previously shown, by relating synchronous production rules to the syntacto-semantic composition process. Our parser achieves an accuracy of 90.35 for EDS (89.51 for DMRS) in terms of elementary dependency match, which is a 4.87 (5.45) point improvement over the best existing data-driven model, indicating, in our view, the importance of linguistically-informed derivation for data-driven semantic parsing. This accuracy is equivalent to that of English Resource Grammar guided models, suggesting that (recurrent) neural network models are able to effectively learn deep linguistic knowledge from annotations.

A DAG automaton is a formal device for manipulating graphs. By augmenting a DAG automaton with transduction rules, a DAG transducer has potential applications in fundamental NLP tasks. In this paper, we propose a novel DAG transducer to perform graph-to-program transformation. The target structure of our transducer is a program licensed by a declarative programming language rather than linguistic structures. By executing such a program, we can easily get a surface string. Our transducer is designed especially for natural language generation (NLG) from type-logical semantic graphs. Taking Elementary Dependency Structures, a format of English Resource Semantics, as input, our NLG system achieves a BLEU-4 score of 68.07. This remarkable result demonstrates the feasibility of applying a DAG transducer to resolve NLG, as well as the effectiveness of our design.

Motivated by the positive impact of empty category on syntactic parsing, we study neural models for pre- and in-parsing detection of empty category, which has not previously been investigated. We find several non-obvious facts: (a) BiLSTM can capture non-local contextual information which is essential for detecting empty categories, (b) even with a BiLSTM, syntactic information is still able to enhance the detection, and (c) automatic detection of empty categories improves parsing quality for overt words. Our neural ECD models outperform the prior state-of-the-art by significant margins.

This paper is concerned with building deep grammatical relation (GR) analysis using data-driven approach. To deal with this problem, we propose graph merging, a new perspective, for building flexible dependency graphs: Constructing complex graphs via constructing simple subgraphs. We discuss two key problems in this perspective: (1) how to decompose a complex graph into simple subgraphs, and (2) how to combine subgraphs into a coherent complex graph. Experiments demonstrate the effectiveness of graph merging. Our parser reaches state-of-the-art performance and is significantly better than two transition-based parsers.

This paper is concerned with whether deep syntactic information can help surface parsing, with a particular focus on empty categories. We design new algorithms to produce dependency trees in which empty elements are allowed, and evaluate the impact of information about empty category on parsing overt elements. Such information is helpful to reduce the approximation error in a structured parsing model, but increases the search space for inference and accordingly the estimation error. To deal with structure-based overfitting, we propose to integrate disambiguation models with and without empty elements, and perform structure regularization via joint decoding. Experiments on English and Chinese TreeBanks with different parsing models indicate that incorporating empty elements consistently improves surface parsing.

We propose a new Maximum Subgraph algorithm for first-order parsing to 1-endpoint-crossing, pagenumber-2 graphs. Our algorithm has two characteristics: (1) it separates the construction for noncrossing edges and crossing edges; (2) in a single construction step, whether to create a new arc is deterministic. These two characteristics make our algorithm relatively easy to be extended to incorporiate crossing-sensitive second-order features. We then introduce a new algorithm for quasi-second-order parsing. Experiments demonstrate that second-order features are helpful for Maximum Subgraph parsing.

We model a dependency graph as a book, a particular kind of topological space, for semantic dependency parsing. The spine of the book is made up of a sequence of words, and each page contains a subset of noncrossing arcs. To build a semantic graph for a given sentence, we design new Maximum Subgraph algorithms to generate noncrossing graphs on each page, and a Lagrangian Relaxation-based algorithm tocombine pages into a book. Experiments demonstrate the effectiveness of the bookembedding framework across a wide range of conditions. Our parser obtains comparable results with a state-of-the-art transition-based parser.

We study the Maximum Subgraph problem in deep dependency parsing. We consider two restrictions to deep dependency graphs: (a) 1-endpoint-crossing and (b) pagenumber-2. Our main contribution is an exact algorithm that obtains maximum subgraphs satisfying both restrictions simultaneously in time O(n5). Moreover, ignoring one linguistically-rare structure descreases the complexity to O(n4). We also extend our quartic-time algorithm into a practical parser with a discriminative disambiguation model and evaluate its performance on four linguistic data sets used in semantic dependency parsing.

We present a comparative study of transition-, graph- and PCFG-based models aimed at illuminating more precisely the likely contribution of CFGs in improving Chinese dependency parsing accuracy, especially by combining heterogeneous models. Inspired by the impact of a constituency grammar on dependency parsing, we propose several strategies to acquire pseudo CFGs only from dependency annotations. Compared to linguistic grammars learned from rich phrase-structure treebanks, well designed pseudo grammars achieve similar parsing accuracy and have equivalent contributions to parser ensemble. Moreover, pseudo grammars increase the diversity of base models; therefore, together with all other models, further improve system combination. Based on automatic POS tagging, our final model achieves a UAS of 87.23%, resulting in a significant improvement of the state of the art.

Automatic acquisition of novel compounds is notoriously difficult because most novel compounds have relatively low frequency in a corpus. The current study proposes a new method to deal with the novel compound acquisition challenge. We model this task as a two-class classification problem in which a candidate compound is either classified as a compound or a non-compound. A machine learning method using SVM, incorporating two types of linguistically motivated features: semantic features and character features, is applied to identify rare but valid noun compounds. We explore two kinds of training data: one is virtual training data which is obtained by three statistical scores, i.e. co-occurrence frequency, mutual information and dependent ratio, from the frequent compounds; the other is real training data which is randomly selected from the infrequent compounds. We conduct comparative experiments, and the experimental results show that even with limited direct evidence in the corpus for the novel compounds, we can make full use of the typical frequent compounds to help in the discovery of the novel compounds.