Large Language Models (LLMs) can generate the same sequences contained in the pre-train corpora, known as memorization.Previous research studied it at a macro level, leaving micro yet important questions under-explored, e.g., what makes sentences memorized, the dynamics when generating memorized sequence, its connection to unmemorized sequence, and its predictability.We answer the above questions by analyzing the relationship of memorization with outputs from LLM, namely, embeddings, probability distributions, and generated tokens.A memorization score is calculated as the overlap between generated tokens and actual continuations when the LLM is prompted with a context sequence from the pre-train corpora.Our findings reveal:(1) The inter-correlation between memorized/unmemorized sentences, model size, continuation size, and context size, as well as the transition dynamics between sentences of different memorization scores,(2) A sudden drop and increase in the frequency of input tokens when generating memorized/unmemorized sequences (boundary effect),(3) Cluster of sentences with different memorization scores in the embedding space,(4) An inverse boundary effect in the entropy of probability distributions for generated memorized/unmemorized sequences,(5) The predictability of memorization is related to model size and continuation length. In addition, we show a Transformer model trained by the hidden states of LLM can predict unmemorized tokens.
Inductive reasoning is fundamental to both human and artificial intelligence. The inductive reasoning abilities of current Large Language Models (LLMs) are evaluated in this research.We argue that only considering induction of rules is too narrow and unrealistic, since inductive reasoning is usually mixed with other abilities, like rules application, results/rules validation, and updated information integration.We probed the LLMs with a set of designed symbolic tasks and found that even state-of-the-art (SotA) LLMs fail significantly, showing the inability of LLMs to perform these intuitively simple tasks.Furthermore, we found that perfect accuracy in a small-size problem does not guarantee the same accuracy in a larger-size version of the same problem, provoking the question of how we can assess the LLMs’ actual problem-solving capabilities.We also argue that Chain-of-Thought prompts help the LLMs by decomposing the problem-solving process, but the LLMs still learn limitedly.Furthermore, we reveal that few-shot examples assist LLM generalization in out-of-domain (OOD) cases, albeit limited. The LLM starts to fail when the problem deviates from the provided few-shot examples.
Unsupervised constituency parsing focuses on identifying word sequences that form a syntactic unit (i.e., constituents) in target sentences. Linguists identify the constituent by evaluating a set of Predicate-Argument Structure (PAS) equivalent sentences where we find the constituent appears more frequently than non-constituents (i.e., the constituent corresponds to a frequent word sequence within the sentence set). However, such frequency information is unavailable in previous parsing methods that identify the constituent by observing sentences with diverse PAS. In this study, we empirically show that constituents correspond to frequent word sequences in the PAS-equivalent sentence set. We propose a frequency-based parser, span-overlap, that (1) computes the span-overlap score as the word sequence’s frequency in the PAS-equivalent sentence set and (2) identifies the constituent structure by finding a constituent tree with the maximum span-overlap score. The parser achieves state-of-the-art level parsing accuracy, outperforming existing unsupervised parsers in eight out of ten languages. Additionally, we discover a multilingual phenomenon: participant-denoting constituents tend to have higher span-overlap scores than equal-length event-denoting constituents, meaning that the former tend to appear more frequently in the PAS-equivalent sentence set than the latter. The phenomenon indicates a statistical difference between the two constituent types, laying the foundation for future labeled unsupervised parsing research.
Recent models in cross-lingual semantic role labeling (SRL) barely analyze the applicability of their network selection.We believe that network selection is important since it affects the transferability of cross-lingual models, i.e., how the model can extract universal features from source languages to label target languages.Therefore, we comprehensively compare the transferability of different graph neural network (GNN)-based models enriched with universal dependency trees.GNN-based models include transformer-based, graph convolutional network-based, and graph attention network (GAT)-based models.We focus our study on a zero-shot setting by training the models in English and evaluating the models in 23 target languages provided by the Universal Proposition Bank.Based on our experiments, we consistently show that syntax from universal dependency trees is essential for cross-lingual SRL models to achieve better transferability.Dependency-aware self-attention with relative position representations (SAN-RPRs) transfer best across languages, especially in the long-range dependency distance.We also show that dependency-aware two-attention relational GATs transfer better than SAN-RPRs in languages where most arguments lie in a 1-2 dependency distance.
This paper focuses on the task of open-domain live commentary generation. Compared to domain-specific work in this task, this setting proved particularly challenging due to the absence of domain-specific features. Aiming to bridge this gap, we integrate spatial information by proposing an utterance generation model with a novel spatial graph that is flexible to deal with the open-domain characteristics of the commentaries and significantly improves performance. Furthermore, we propose a novel evaluation scheme, more suitable for live commentary generation, that uses LLMs to automatically check whether generated utterances address essential aspects of the video via the answerability of questions extracted directly from the videos using LVLMs. Our results suggest that using a combination of our answerability score and a standard machine translation metric is likely a more reliable way to evaluate the performance in this task.
When engaging in conversations, dialogue agents in a virtual simulation environment may exhibit their own emotional states that are unrelated to the immediate conversational context, a phenomenon known as self-emotion. This study explores how such self-emotion affects the agents’ behaviors in dialogue strategies and decision-making within a large language model (LLM)-driven simulation framework. In a dialogue strategy prediction experiment, we analyze the dialogue strategy choices employed by agents both with and without self-emotion, comparing them to those of humans. The results show that incorporating self-emotion helps agents exhibit more human-like dialogue strategies. In an independent experiment comparing the performance of models fine-tuned on GPT-4 generated dialogue datasets, we demonstrate that self-emotion can lead to better overall naturalness and humanness. Finally, in a virtual simulation environment where agents have free discussions, we show that self-emotion of agents can significantly influence the decision-making process of the agents, leading to approximately a 50% change in decisions.
This paper aims to forecast the implicit emotion elicited in the dialogue partner by a textual input utterance. Forecasting the interlocutor’s emotion is beneficial for natural language generation in dialogue systems to avoid generating utterances that make the users uncomfortable. Previous studies forecast the emotion conveyed in the interlocutor’s response, assuming it will explicitly reflect their elicited emotion. However, true emotions are not always expressed verbally. We propose a new task to directly forecast the implicit emotion elicited by an input utterance, which does not rely on this assumption. We compare this task with related ones to investigate the impact of dialogue history and one’s own utterance on predicting explicit and implicit emotions. Our result highlights the importance of dialogue history for predicting implicit emotions. It also reveals that, unlike explicit emotions, implicit emotions show limited improvement in predictive performance with one’s own utterance, and that they are more difficult to predict than explicit emotions. We find that even a large language model (LLM) struggles to forecast implicit emotions accurately.
We propose a method that extends a BART-based language generator using a plug-and-play model to control the rhetorical structure of generated text. Our approach considers rhetorical relations between clauses and generates sentences that reflect this structure using plug-and-play language models. We evaluated our method using the Newsela corpus, which consists of texts at various levels of English proficiency. Our experiments demonstrated that our method outperforms the vanilla BART in terms of the correctness of output discourse and rhetorical structures. In existing methods, the rhetorical structure tends to deteriorate when compared to the baseline, the vanilla BART, as measured by n-gram overlap metrics such as BLEU. However, our proposed method does not exhibit this significant deterioration, demonstrating its advantage.
Previous methods based on Large Language Models (LLM) perform unsupervised dependency parsing by maximizing bi-lexical dependence scores. However, these previous methods adopt dependence scores that are difficult to interpret. These methods cannot incorporate grammatical constraints that previous grammar-based parsing research has shown beneficial to improving parsing performance. In this work, we apply Conditional Mutual Information (CMI), an interpretable metric, to measure the bi-lexical dependence and incorporate grammatical constraints into LLM-based unsupervised parsing. We incorporate Part-Of-Speech information as a grammatical constraint at the CMI estimation stage and integrate two additional grammatical constraints at the subsequent tree decoding stage. We find that the CMI score positively correlates with syntactic dependencies and has a stronger correlation with the syntactic dependencies than baseline scores. Our experiment confirms the benefits and applicability of the proposed grammatical constraints across five languages and eight datasets. The CMI parsing model outperforms state-of-the-art LLM-based models and similarly constrained grammar-based models. Our analysis reveals that the CMI model is strong in retrieving dependency relations with rich lexical interactions but is weak in retrieving relations with sparse lexical interactions, indicating a potential limitation in CMI-based unsupervised parsing methods.
This paper critically examines the arithmetic capabilities of Large Language Models (LLMs), uncovering significant limitations in their performance. Our research reveals a notable decline in accuracy for complex calculations involving large numbers, with addition and subtraction tasks showing varying degrees of proficiency. Additionally, we challenge the notion that arithmetic is language-independent, finding up to a 10% difference in performance across twenty languages. The study also compares self-verification methods with cross-agent collaborations, showing that a single model often outperforms collaborative approaches in basic arithmetic tasks. These findings suggest a need to reassess the effectiveness of LLMs in tasks requiring numerical accuracy and precision.
We investigate intention detection in persuasive multi-turn dialogs employing the largest available Large Language Models (LLMs).Much of the prior research measures the intention detection capability of machine learning models without considering the conversational history.To evaluate LLMs’ intention detection capability in conversation, we modified the existing datasets of persuasive conversation and created datasets using a multiple-choice paradigm.It is crucial to consider others’ perspectives through their utterances when engaging in a persuasive conversation, especially when making a request or reply that is inconvenient for others.This feature makes the persuasive dialogue suitable for the dataset of measuring intention detection capability.We incorporate the concept of ‘face acts,’ which categorize how utterances affect mental states.This approach enables us to measure intention detection capability by focusing on crucial intentions and to conduct comprehensible analysis according to intention types.
Previously, we introduced a method to generate a multilingual Combinatory Categorial Grammar (CCG) treebank by converting from the Universal Dependencies (UD). However, the method only produces bare CCG derivations without any accompanying semantic representations, which makes it difficult to obtain satisfactory analyses for constructions that involve non-local dependencies, such as control/raising or relative clauses, and limits the general applicability of the treebank. In this work, we present an algorithm that adds semantic representations to existing CCG derivations, in the form of predicate-argument structures. Through hand-crafted rules, we enhance each CCG category with headedness information, with which both local and non-local dependencies can be properly projected. This information is extracted from various sources, including UD, Enhanced UD, and proposition banks. Evaluation of our projected dependencies on the English PropBank and the Universal PropBank 2.0 shows that they can capture most of the semantic dependencies in the target corpora. Further error analysis measures the effectiveness of our algorithm for each language tested, and reveals several issues with the previous method and source data.
In automated scientific document analysis, accurately interpreting math formulae is imperative alongside comprehending natural language. Ambiguity in math identifiers within a single document poses significant challenges to understanding math formulae. While disambiguating math identifiers across documents has seen some progress, resolving ambiguity within a document remains inadequately researched due to complexity and insufficient datasets. The level of difficulty and information required to accomplish this task was uncertain. This study aims to determine which information is necessary for the intra-document disambiguation of math identifiers. Our findings indicate that the position data and local formula structure surrounding the identifiers, including modifiers, are particularly critical. For our study, we expanded a dataset for formula grounding and doubled its size to include annotations for 27,655 math identifier occurrences. We have created a multi-layer perceptron model that performs similarly to humans, with an 85% accuracy and a kappa value of 0.73, outperforming rule-based baselines. We trained and evaluated the model with papers in natural language processing (NLP). Our findings were also confirmed valid in fields other than NLP by applying the trained models to papers from various fields. These results will aid in improving mathematical language processing, such as mathematical information retrieval.
The task of quote attribution seeks to pair textual utterances with the name of their speakers. Despite continuing research efforts on the task, models are rarely evaluated systematically against previous models in comparable settings on the same datasets. This has resulted in a poor understanding of the relative strengths and weaknesses of various approaches. In this work we formalize the task of quote attribution, and in doing so, establish a basis of comparison across existing models. We present an exhaustive benchmark of known models, including natural extensions to larger LLM base models, on all available datasets in both English and Chinese. Our benchmarking results reveal that the CEQA model attains state-of-the-art performance among all supervised methods, and ChatGPT, operating in a four-shot setting, demonstrates performance on par with or surpassing that of supervised methods on some datasets. Detailed error analysis identify several key factors contributing to prediction errors.
This paper presents the formalization of tree-shape uncertainty that enables us to analyze the inherent branching bias of unsupervised parsing models using raw texts alone. Previous work analyzed the branching bias of unsupervised parsing models by comparing the outputs of trained parsers with gold syntactic trees. However, such approaches do not consider the fact that texts can be generated by different grammars with different syntactic trees, possibly failing to clearly separate the inherent bias of the model and the bias in train data learned by the model. To this end, we formulate tree-shape uncertainty and derive sufficient conditions that can be used for creating texts that are expected to contain no biased information on branching. In the experiment, we show that training parsers on such unbiased texts can effectively detect the branching bias of existing unsupervised parsing models. Such bias may depend only on the algorithm, or it may depend on seemingly unrelated dataset statistics such as sequence length and vocabulary size.
Numbers have unique characteristics to words. Teaching models to understand numbers in text is an open-ended research question. Instead of discussing the required calculation skills, this paper focuses on a more fundamental topic: understanding numerals. We point out that innumeracy—the inability to handle basic numeral concepts—exists in most pretrained language models (LMs), and we propose a method to solve this issue by exploring the notation of numbers. Further, we discuss whether changing notation and pre-finetuning along with the comparing-number task can improve performance in three benchmark datasets containing quantitative-related tasks. The results of this study indicate that input reframing and the proposed pre-finetuning task is useful for RoBERTa.
Existing dialogue models may encounter scenarios which are not well-represented in the training data, and as a result generate responses that are unnatural, inappropriate, or unhelpful. We propose the “Ask an Expert” framework in which the model is trained with access to an “expert” which it can consult at each turn. Advice is solicited via a structured dialogue with the expert, and the model is optimized to selectively utilize (or ignore) it given the context and dialogue history. In this work the expert takes the form of an LLM.We evaluate this framework in a mental health support domain, where the structure of the expert conversation is outlined by pre-specified prompts which reflect a reasoning strategy taught to practitioners in the field. Blenderbot models utilizing “Ask an Expert” show quality improvements across all expert sizes, including those with fewer parameters than the dialogue model itself. Our best model provides a ~10% improvement over baselines, approaching human-level scores on “engingingness” and “helpfulness” metrics.
This paper explores the task of Temporal Video Grounding (TVG) where, given an untrimmed video and a query sentence, the goal is to recognize and determine temporal boundaries of action instances in the video described by natural language queries. Recent works tackled this task by improving query inputs with large pre-trained language models (PLM), at the cost of more expensive training. However, the effects of this integration are unclear, as these works also propose improvements in the visual inputs. Therefore, this paper studies the role of query sentence representation with PLMs in TVG and assesses the applicability of parameter-efficient training with NLP adapters. We couple popular PLMs with a selection of existing approaches and test different adapters to reduce the impact of the additional parameters. Our results on three challenging datasets show that, with the same visual inputs, TVG models greatly benefited from the PLM integration and fine-tuning, stressing the importance of the text query representation in this task. Furthermore, adapters were an effective alternative to full fine-tuning, even though they are not tailored to our task, allowing PLM integration in larger TVG models and delivering results comparable to SOTA models. Finally, our results shed light on which adapters work best in different scenarios.
Knowing how to end and resume conversations over time is a natural part of communication, allowing for discussions to span weeks, months, or years. The duration of gaps between conversations dictates which topics are relevant and which questions to ask, and dialogue systems which do not explicitly model time may generate responses that are unnatural. In this work we explore the idea of making dialogue models aware of time, and present GapChat, a multi-session dialogue dataset in which the time between each session varies. While the dataset is constructed in real-time, progress on events in speakers’ lives is simulated in order to create realistic dialogues occurring across a long timespan. We expose time information to the model and compare different representations of time and event progress. In human evaluation we show that time-aware models perform better in metrics that judge the relevance of the chosen topics and the information gained from the conversation.
We explore the idea of incorporating concepts from writing skills curricula into human-machine collaborative writing scenarios, focusing on adding writing modes as a control for text generation models. Using crowd-sourced workers, we annotate a corpus of narrative text paragraphs with writing mode labels. Classifiers trained on this data achieve an average accuracy of ~87% on held-out data. We fine-tune a set of large language models to condition on writing mode labels, and show that the generated text is recognized as belonging to the specified mode with high accuracy. To study the ability of writing modes to provide fine-grained control over generated text, we devise a novel turn-based text reconstruction game to evaluate the difference between the generated text and the author’s intention. We show that authors prefer text suggestions made by writing mode-controlled models on average 61.1% of the time, with satisfaction scores 0.5 higher on a 5-point ordinal scale. When evaluated by humans, stories generated via collaboration with writing mode-controlled models achieve high similarity with the professionally written target story. We conclude by identifying the most common mistakes found in the generated stories.
Live commentaries are essential for enhancing spectators’ enjoyment and understanding during sports events or e-sports streams. We introduce a live audio commentator system designed specifically for a racing game, driven by the high demand in the e-sports field. While a player is playing a racing game, our system tracks real-time user play data including speed and steer rotations, and generates commentary to accompany the live stream. Human evaluation suggested that generated commentary enhances enjoyment and understanding of races compared to streams without commentary. Incorporating additional modules to improve diversity and detect irregular events, such as course-outs and collisions, further increases the preference for the output commentaries.
In this paper, we propose a mixture model-based end-to-end method to model the syntactic-semantic dependency correlation in Semantic Role Labeling (SRL). Semantic dependencies in SRL are modeled as a distribution over semantic dependency labels conditioned on a predicate and an argument word. The semantic label distribution varies depending on Shortest Syntactic Dependency Path (SSDP) hop patterns. We target the variation of semantic label distributions using a mixture model, separately estimating semantic label distributions for different hop patterns and probabilistically clustering hop patterns with similar semantic label distributions. Experiments show that the proposed method successfully learns a cluster assignment reflecting the variation of semantic label distributions. Modeling the variation improves performance in predicting short distance semantic dependencies, in addition to the improvement on long distance semantic dependencies that previous syntax-aware methods have achieved. The proposed method achieves a small but statistically significant improvement over baseline methods in English, German, and Spanish and obtains competitive performance with state-of-the-art methods in English.
Existing automatic story evaluation methods place a premium on story lexical level coherence, deviating from human preference.We go beyond this limitation by considering a novel Story Evaluation method that mimics human preference when judging a story, namely StoryER, which consists of three sub-tasks: Ranking, Rating and Reasoning.Given either a machine-generated or a human-written story, StoryER requires the machine to output 1) a preference score that corresponds to human preference, 2) specific ratings and their corresponding confidences and 3) comments for various aspects (e.g., opening, character-shaping).To support these tasks, we introduce a well-annotated dataset comprising (i) 100k ranked story pairs; and (ii) a set of 46k ratings and comments on various aspects of the story.We finetune Longformer-Encoder-Decoder (LED) on the collected dataset, with the encoder responsible for preference score and aspect prediction and the decoder for comment generation.Our comprehensive experiments result a competitive benchmark for each task, showing the high correlation to human preference.In addition, we have witnessed the joint learning of the preference scores, the aspect ratings, and the comments brings gain each single task.Our dataset and benchmarks are publicly available to advance the research of story evaluation tasks.
Live commentary plays an important role in sports broadcasts and video games, making spectators more excited and immersed. In this context, though approaches for automatically generating such commentary have been proposed in the past, they have been generally concerned with specific fields, where it is possible to leverage domain-specific information. In light of this, we propose the task of generating video commentary in an open-domain fashion. We detail the construction of a new large-scale dataset of transcribed commentary aligned with videos containing various human actions in a variety of domains, and propose approaches based on well-known neural architectures to tackle the task. To understand the strengths and limitations of current approaches, we present an in-depth empirical study based on our data. Our results suggest clear trade-offs between textual and visual inputs for the models and highlight the importance of relying on external knowledge in this open-domain setting, resulting in a set of robust baselines for our task.
We introduce the task of implicit offensive text detection in dialogues, where a statement may have either an offensive or non-offensive interpretation, depending on the listener and context. We argue that reasoning is crucial for understanding this broader class of offensive utterances, and release SLIGHT, a dataset to support research on this task. Experiments using the data show that state-of-the-art methods of offense detection perform poorly when asked to detect implicitly offensive statements, achieving only ∼ 11% accuracy. In contrast to existing offensive text detection datasets, SLIGHT features human-annotated chains of reasoning which describe the mental process by which an offensive interpretation can be reached from each ambiguous statement. We explore the potential for a multi-hop reasoning approach by utilizing existing entailment models to score the probability of these chains, and show that even naive reasoning models can yield improved performance in most situations. Analysis of the chains provides insight into the human interpretation process and emphasizes the importance of incorporating additional commonsense knowledge.
This paper proposes a data representation framework for semantic parsing and task-oriented dialogue systems, aiming to achieve a uniform representation for syntactically and semantically diverse machine-readable formats.Current NLP systems heavily rely on adapting pre-trained language models to specific tasks, and this approach has been proven effective for modeling natural language texts.However, little attention has been paid to the representation of machine-readable formats, such as database queries and dialogue states.We present a method for converting original machine-readable formats of semantic parsing and task-oriented dialogue datasets into a syntactically and semantically uniform representation.We define a meta grammar for syntactically uniform representations and translate semantically equivalent functions into a uniform vocabulary.Empirical experiments on 13 datasets show that accuracy consistently improves over original formats, revealing the advantage of the proposed representation.Additionally, we show that the proposed representation allows for transfer learning across datasets.
In this paper, we specifically look at the image-text retrieval problem. Recent multimodal frameworks have shown that structured inputs and fine-tuning lead to consistent performance improvement. However, this paradigm has been challenged recently with newer Transformer-based models that can reach zero-shot state-of-the-art results despite not explicitly using structured data during pre-training. Since such strategies lead to increased computational resources, we seek to better understand their role in image-text retrieval by analyzing visual and text representations extracted with three multimodal frameworks – SGM, UNITER, and CLIP. To perform such analysis, we represent a single image or text as low-dimensional linear subspaces and perform retrieval based on subspace similarity. We chose this representation as subspaces give us the flexibility to model an entity based on feature sets, allowing us to observe how integrating or reducing information changes the representation of each entity. We analyze the performance of the selected models’ features on two standard benchmark datasets. Our results indicate that heavily pre-training models can already lead to features with critical information representing each entity, with zero-shot UNITER features performing consistently better than fine-tuned features. Furthermore, while models can benefit from structured inputs, learning representations for objects and relationships separately, such as in SGM, likely causes a loss of crucial contextual information needed to obtain a compact cluster that can effectively represent a single entity.
Incorporating stronger syntactic biases into neural language models (LMs) is a long-standing goal, but research in this area often focuses on modeling English text, where constituent treebanks are readily available. Extending constituent tree-based LMs to the multilingual setting, where dependency treebanks are more common, is possible via dependency-to-constituency conversion methods. However, this raises the question of which tree formats are best for learning the model, and for which languages. We investigate this question by training recurrent neural network grammars (RNNGs) using various conversion methods, and evaluating them empirically in a multilingual setting. We examine the effect on LM performance across nine conversion methods and five languages through seven types of syntactic tests. On average, the performance of our best model represents a 19 % increase in accuracy over the worst choice across all languages. Our best model shows the advantage over sequential/overparameterized LMs, suggesting the positive effect of syntax injection in a multilingual setting. Our experiments highlight the importance of choosing the right tree formalism, and provide insights into making an informed decision.
Grounding the meaning of each symbol in math formulae is important for automated understanding of scientific documents. Generally speaking, the meanings of math symbols are not necessarily constant, and the same symbol is used in multiple meanings. Therefore, coreference relations between symbols need to be identified for grounding, and the task has aspects of both description alignment and coreference analysis. In this study, we annotated 15 papers selected from arXiv.org with the grounding information. In total, 12,352 occurrences of math identifiers in these papers were annotated, and all coreference relations between them were made explicit in each paper. The constructed dataset shows that regardless of the ambiguity of symbols in math formulae, coreference relations can be labeled with a high inter-annotator agreement. The constructed dataset enables us to achieve automation of formula grounding, and in turn, make deeper use of the knowledge in scientific documents using techniques such as math information extraction. The built grounding dataset is available at https://sigmathling.kwarc.info/resources/grounding- dataset/.
This paper introduces an algorithm to convert Universal Dependencies (UD) treebanks to Combinatory Categorial Grammar (CCG) treebanks. As CCG encodes almost all grammatical information into the lexicon, obtaining a high-quality CCG derivation from a dependency tree is a challenging task. Our algorithm relies on hand-crafted rules to assign categories to constituents, and a non-statistical parser to derive full CCG parses given the assigned categories. To evaluate our converted treebanks, we perform lexical, sentential, and syntactic rule coverage analysis, as well as CCG parsing experiments. Finally, we discuss how our method handles complex constructions, and propose possible future extensions.
When individuals communicate with each other, they use different vocabulary, speaking speed, facial expressions, and body language depending on the people they talk to. This paper focuses on the speaker’s age as a factor that affects the change in communication. We collected a multimodal dialogue corpus with a wide range of speaker ages. As a dialogue task, we focus on travel, which interests people of all ages, and we set up a task based on a tourism consultation between an operator and a customer at a travel agency. This paper provides details of the dialogue task, the collection procedure and annotations, and the analysis on the characteristics of the dialogues and facial expressions focusing on the age of the speakers. Results of the analysis suggest that the adult speakers have more independent opinions, the older speakers more frequently express their opinions frequently compared with other age groups, and the operators expressed a smile more frequently to the minor speakers.
Controlling the generation of image captions attracts lots of attention recently. In this paper, we propose a framework leveraging partial syntactic dependency trees as control signals to make image captions include specified words and their syntactic structures. To achieve this purpose, we propose a Syntactic Dependency Structure Aware Model (SDSAM), which explicitly learns to generate the syntactic structures of image captions to include given partial dependency trees. In addition, we come up with a metric to evaluate how many specified words and their syntactic dependencies are included in generated captions. We carry out experiments on two standard datasets: Microsoft COCO and Flickr30k. Empirical results show that image captions generated by our model are effectively controlled in terms of specified words and their syntactic structures. The code is available on GitHub.
We propose a methodology for representing the reasoning structure of arguments using Bayesian networks and predicate logic facilitated by argumentation schemes. We express the meaning of text segments using predicate logic and map the boolean values of predicate logic expressions to nodes in a Bayesian network. The reasoning structure among text segments is described with a directed acyclic graph. While our formalism is highly expressive and capable of describing the informal logic of human arguments, it is too open-ended to actually build a network for an argument. It is not at all obvious which segment of argumentative text should be considered as a node in a Bayesian network, and how to decide the dependencies among nodes. To alleviate the difficulty, we provide abstract network fragments, called idioms, which represent typical argument justification patterns derived from argumentation schemes. The network construction process is decomposed into idiom selection, idiom instantiation, and idiom combination. We define 17 idioms in total by referring to argumentation schemes as well as analyzing actual arguments and fitting idioms to them. We also create a dataset consisting of pairs of an argumentative text and a corresponding Bayesian network. Our dataset contains about 2,400 pairs, which is large in the research area of argumentation schemes.
We propose the task of automatically generating commentaries for races in a motor racing game, from vision, structured numerical, and textual data. Commentaries provide information to support spectators in understanding events in races. Commentary generation models need to interpret the race situation and generate the correct content at the right moment. We divide the task into two subtasks: utterance timing identification and utterance generation. Because existing datasets do not have such alignments of data in multiple modalities, this setting has not been explored in depth. In this study, we introduce a new large-scale dataset that contains aligned video data, structured numerical data, and transcribed commentaries that consist of 129,226 utterances in 1,389 races in a game. Our analysis reveals that the characteristics of commentaries change over time or from viewpoints. Our experiments on the subtasks show that it is still challenging for a state-of-the-art vision encoder to capture useful information from videos to generate accurate commentaries. We make the dataset and baseline implementation publicly available for further research.
A large amount of scientific knowledge is represented within mixed forms of natural language texts and mathematical formulae. Therefore, a collaboration of natural language processing and formula analyses, so-called mathematical language processing, is necessary to enable computers to understand and retrieve information from the documents. However, as we will show in this project, a mathematical notation can change its meaning even within the scope of a single paragraph. This flexibility makes it difficult to extract the exact meaning of a mathematical formula. In this project, we will propose a new task direction for grounding mathematical formulae. Particularly, we are addressing the widespread misconception of various research projects in mathematical information retrieval, which presume that mathematical notations have a fixed meaning within a single document. We manually annotated a long scientific paper to illustrate the task concept. Our high inter-annotator agreement shows that the task is well understood for humans. Our results indicate that it is worthwhile to grow the techniques for the proposed task to contribute to the further progress of mathematical language processing.
Existing models for data-to-text tasks generate fluent but sometimes incorrect sentences e.g., “Nikkei gains” is generated when “Nikkei drops” is expected. We investigate models trained on contrastive examples i.e., incorrect sentences or terms, in addition to correct ones to reduce such errors. We first create rules to produce contrastive examples from correct ones by replacing frequent crucial terms such as “gain” or “drop”. We then use learning methods with several losses that exploit contrastive examples. Experiments on the market comment generation task show that 1) exploiting contrastive examples improves the capability of generating sentences with better lexical choice, without degrading the fluency, 2) the choice of the loss function is an important factor because the performances on different metrics depend on the types of loss functions, and 3) the use of the examples produced by some specific rules further improves performance. Human evaluation also supports the effectiveness of using contrastive examples.
In this paper, we evaluate the progress of our field toward solving simple factoid questions over a knowledge base, a practically important problem in natural language interface to database. As in other natural language understanding tasks, a common practice for this task is to train and evaluate a model on a single dataset, and recent studies suggest that SimpleQuestions, the most popular and largest dataset, is nearly solved under this setting. However, this common setting does not evaluate the robustness of the systems outside of the distribution of the used training data. We rigorously evaluate such robustness of existing systems using different datasets. Our analysis, including shifting of training and test datasets and training on a union of the datasets, suggests that our progress in solving SimpleQuestions dataset does not indicate the success of more general simple question answering. We discuss a possible future direction toward this goal.
Many researchers have tried to predict the accuracies of extrinsic evaluation by using intrinsic evaluation to evaluate word embedding. The relationship between intrinsic and extrinsic evaluation, however, has only been studied with simple correlation analysis, which has difficulty capturing complex cause-effect relationships and integrating external factors such as the hyperparameters of word embedding. To tackle this problem, we employ partial least squares path modeling (PLS-PM), a method of structural equation modeling developed for causal analysis. We propose a causal diagram consisting of the evaluation results on the BATS, VecEval, and SentEval datasets, with a causal hypothesis that linguistic knowledge encoded in word embedding contributes to solving downstream tasks. Our PLS-PM models are estimated with 600 word embeddings, and we prove the existence of causal relations between linguistic knowledge evaluated on BATS and the accuracies of downstream tasks evaluated on VecEval and SentEval in our PLS-PM models. Moreover, we show that the PLS-PM models are useful for analyzing the effect of hyperparameters, including the training algorithm, corpus, dimension, and context window, and for validating the effectiveness of intrinsic evaluation.
End-to-end models on data-to-text learn the mapping of data and text from the aligned pairs in the dataset. However, these alignments are not always obtained reliably, especially for the time-series data, for which real time comments are given to some situation and there might be a delay in the comment delivery time compared to the actual event time. To handle this issue of possible noisy alignments in the dataset, we propose a neural network model with multi-timestep data and a copy mechanism, which allows the models to learn the correspondences between data and text from the dataset with noisier alignments. We focus on generating market comments in Japanese that are delivered each time an event occurs in the market. The core idea of our approach is to utilize multi-timestep data, which is not only the latest market price data when the comment is delivered, but also the data obtained at several timesteps earlier. On top of this, we employ a copy mechanism that is suitable for referring to the content of data records in the market price data. We confirm the superiority of our proposal by two evaluation metrics and show the accuracy improvement of the sentence generation using the time series data by our proposed method.
This paper describes a method for annotating the Japanese Sign Language (JSL) dialogue corpus. We developed a way to identify interactional boundaries and define a ‘utterance unit’ in sign language using various multimodal features accompanying signing. The utterance unit is an original concept for segmenting and annotating sign language dialogue referring to signer’s native sense from the perspectives of Conversation Analysis (CA) and Interaction Studies. First of all, we postulated that we should identify a fundamental concept of interaction-specific unit for understanding interactional mechanisms, such as turn-taking (Sacks et al. 1974), in sign-language social interactions. Obviously, it does should not relying on a spoken language writing system for storing signings in corpora and making translations. We believe that there are two kinds of possible applications for utterance units: one is to develop corpus linguistics research for both signed and spoken corpora; the other is to build an informatics system that includes, but is not limited to, a machine translation system for sign languages.
The global pandemic of COVID-19 has made the public pay close attention to related news, covering various domains, such as sanitation, treatment, and effects on education. Meanwhile, the COVID-19 condition is very different among the countries (e.g., policies and development of the epidemic), and thus citizens would be interested in news in foreign countries. We build a system for worldwide COVID-19 information aggregation containing reliable articles from 10 regions in 7 languages sorted by topics. Our reliable COVID-19 related website dataset collected through crowdsourcing ensures the quality of the articles. A neural machine translation module translates articles in other languages into Japanese and English. A BERT-based topic-classifier trained on our article-topic pair dataset helps users find their interested information efficiently by putting articles into different categories.
In this paper, we compare four state-of-the-art neural network dependency parsers for the Semitic language Amharic. As Amharic is a morphologically-rich and less-resourced language, the out-of-vocabulary (OOV) problem will be higher when we develop data-driven models. This fact limits researchers to develop neural network parsers because the neural network requires large quantities of data to train a model. We empirically evaluate neural network parsers when a small Amharic treebank is used for training. Based on our experiment, we obtain an 83.79 LAS score using the UDPipe system. Better accuracy is achieved when the neural parsing system uses external resources like word embedding. Using such resources, the LAS score for UDPipe improves to 85.26. Our experiment shows that the neural networks can learn dependency relations better from limited data while segmentation and POS tagging require much data.
We propose a data-to-text generation model with two modules, one for tracking and the other for text generation. Our tracking module selects and keeps track of salient information and memorizes which record has been mentioned. Our generation module generates a summary conditioned on the state of tracking module. Our proposed model is considered to simulate the human-like writing process that gradually selects the information by determining the intermediate variables while writing the summary. In addition, we also explore the effectiveness of the writer information for generations. Experimental results show that our proposed model outperforms existing models in all evaluation metrics even without writer information. Incorporating writer information further improves the performance, contributing to content planning and surface realization.
Peer review is a core element of the scientific process, particularly in conference-centered fields such as ML and NLP. However, only few studies have evaluated its properties empirically. Aiming to fill this gap, we present a corpus that contains over 4k reviews and 1.2k author responses from ACL-2018. We quantitatively and qualitatively assess the corpus. This includes a pilot study on paper weaknesses given by reviewers and on quality of author responses. We then focus on the role of the rebuttal phase, and propose a novel task to predict after-rebuttal (i.e., final) scores from initial reviews and author responses. Although author responses do have a marginal (and statistically significant) influence on the final scores, especially for borderline papers, our results suggest that a reviewer’s final score is largely determined by her initial score and the distance to the other reviewers’ initial scores. In this context, we discuss the conformity bias inherent to peer reviewing, a bias that has largely been overlooked in previous research. We hope our analyses will help better assess the usefulness of the rebuttal phase in NLP conferences.
We propose a data-to-document generator that can easily control the contents of output texts based on a neural language model. Conventional data-to-text model is useful when a reader seeks a global summary of data because it has only to describe an important part that has been extracted beforehand. However, because depending on users, it differs what they are interested in, so it is necessary to develop a method to generate various summaries according to users’ interests. We develop a model to generate various summaries and to control their contents by providing the explicit targets for a reference to the model as controllable factors. In the experiments, we used five-minute or one-hour charts of 9 indicators (e.g., Nikkei225), as time-series data, and daily summaries of Nikkei Quick News as textual data. We conducted comparative experiments using two pieces of information: human-designed topic labels indicating the contents of a sentence and automatically extracted keywords as the referential information for generation.
Recognizing temporal relations among events and time expressions has been an essential but challenging task in natural language processing. Conventional annotation of judging temporal relations puts a heavy load on annotators. In reality, the existing annotated corpora include annotations on only “salient” event pairs, or on pairs in a fixed window of sentences. In this paper, we propose a new approach to obtain temporal relations from absolute time value (a.k.a. time anchors), which is suitable for texts containing rich temporal information such as news articles. We start from time anchors for events and time expressions, and temporal relation annotations are induced automatically by computing relative order of two time anchors. This proposal shows several advantages over the current methods for temporal relation annotation: it requires less annotation effort, can induce inter-sentence relations easily, and increases informativeness of temporal relations. We compare the empirical statistics and automatic recognition results with our data against a previous temporal relation corpus. We also reveal that our data contributes to a significant improvement of the downstream time anchor prediction task, demonstrating 14.1 point increase in overall accuracy.
Syntactic parsing plays a crucial role in improving the quality of natural language processing tasks. Although there have been several research projects on syntactic parsing in Vietnamese, the parsing quality has been far inferior than those reported in major languages, such as English and Chinese. In this work, we evaluated representative constituency parsing models on a Vietnamese Treebank to look for the most suitable parsing method for Vietnamese. We then combined the advantages of automatic and manual analysis to investigate errors produced by the experimented parsers and find the reasons for them. Our analysis focused on three possible sources of parsing errors, namely limited training data, part-of-speech (POS) tagging errors, and ambiguous constructions. As a result, we found that the last two sources, which frequently appear in Vietnamese text, significantly attributed to the poor performance of Vietnamese parsing.
This paper discusses the representation of coordinate structures in the Universal Dependencies framework for two head-final languages, Japanese and Korean. UD applies a strict principle that makes the head of coordination the left-most conjunct. However, the guideline may produce syntactic trees which are difficult to accept in head-final languages. This paper describes the status in the current Japanese and Korean corpora and proposes alternative designs suitable for these languages.
Comments on a stock market often include the reason or cause of changes in stock prices, such as “Nikkei turns lower as yen’s rise hits exporters.” Generating such informative sentences requires capturing the relationship between different resources, including a target stock price. In this paper, we propose a model for automatically generating such informative market comments that refer to external resources. We evaluated our model through an automatic metric in terms of BLEU and human evaluation done by an expert in finance. The results show that our model outperforms the existing model both in BLEU scores and human judgment.
We approach the recognition of textual entailment using logical semantic representations and a theorem prover. In this setup, lexical divergences that preserve semantic entailment between the source and target texts need to be explicitly stated. However, recognising subsentential semantic relations is not trivial. We address this problem by monitoring the proof of the theorem and detecting unprovable sub-goals that share predicate arguments with logical premises. If a linguistic relation exists, then an appropriate axiom is constructed on-demand and the theorem proving continues. Experiments show that this approach is effective and precise, producing a system that outperforms other logic-based systems and is competitive with state-of-the-art statistical methods.
This paper presents a novel encoder-decoder model for automatically generating market comments from stock prices. The model first encodes both short- and long-term series of stock prices so that it can mention short- and long-term changes in stock prices. In the decoding phase, our model can also generate a numerical value by selecting an appropriate arithmetic operation such as subtraction or rounding, and applying it to the input stock prices. Empirical experiments show that our best model generates market comments at the fluency and the informativeness approaching human-generated reference texts.
Temporal relation classification is becoming an active research field. Lots of methods have been proposed, while most of them focus on extracting features from external resources. Less attention has been paid to a significant advance in a closely related task: relation extraction. In this work, we borrow a state-of-the-art method in relation extraction by adopting bidirectional long short-term memory (Bi-LSTM) along dependency paths (DP). We make a “common root” assumption to extend DP representations of cross-sentence links. In the final comparison to two state-of-the-art systems on TimeBank-Dense, our model achieves comparable performance, without using external knowledge, as well as manually annotated attributes of entities (class, tense, polarity, etc.).
Treebanks are important resources for researchers in natural language processing, speech recognition, theoretical linguistics, etc. To strengthen the automatic processing of the Vietnamese language, a Vietnamese treebank has been built. However, the quality of this treebank is not satisfactory and is a possible source for the low performance of Vietnamese language processing. We have been building a new treebank for Vietnamese with about 40,000 sentences annotated with three layers: word segmentation, part-of-speech tagging, and bracketing. In this paper, we describe several challenges of Vietnamese language and how we solve them in developing annotation guidelines. We also present our methods to improve the quality of the annotation guidelines and ensure annotation accuracy and consistency. Experiment results show that inter-annotator agreement ratios and accuracy are higher than 90% which is satisfactory.
We present an attempt to port the international syntactic annotation scheme, Universal Dependencies, to the Japanese language in this paper. Since the Japanese syntactic structure is usually annotated on the basis of unique chunk-based dependencies, we first introduce word-based dependencies by using a word unit called the Short Unit Word, which usually corresponds to an entry in the lexicon UniDic. Porting is done by mapping the part-of-speech tagset in UniDic to the universal part-of-speech tagset, and converting a constituent-based treebank to a typed dependency tree. The conversion is not straightforward, and we discuss the problems that arose in the conversion and the current solutions. A treebank consisting of 10,000 sentences was built by converting the existent resources and currently released to the public.
We describe our ongoing effort to establish an annotation scheme for describing the semantic structures of research articles in the computer science domain, with the intended use of developing search systems that can refine their results by the roles of the entities denoted by the query keys. In our scheme, mentions of entities are annotated with ontology-based types, and the roles of the entities are annotated as relations with other entities described in the text. So far, we have annotated 400 abstracts from the ACL anthology and the ACM digital library. In this paper, the scheme and the annotated dataset are described, along with the problems found in the course of annotation. We also show the results of automatic annotation and evaluate the corpus in a practical setting in application to topic extraction.
We announce a new language resource for research on semantic parsing, a large, carefully curated collection of semantic dependency graphs representing multiple linguistic traditions. This resource is called SDP~2016 and provides an update and extension to previous versions used as Semantic Dependency Parsing target representations in the 2014 and 2015 Semantic Evaluation Exercises. For a common core of English text, this third edition comprises semantic dependency graphs from four distinct frameworks, packaged in a unified abstract format and aligned at the sentence and token levels. SDP 2016 is the first general release of this resource and available for licensing from the Linguistic Data Consortium in May 2016. The data is accompanied by an open-source SDP utility toolkit and system results from previous contrastive parsing evaluations against these target representations.
Automatic video description generation has recently been getting attention after rapid advancement in image caption generation. Automatically generating description for a video is more challenging than for an image due to its temporal dynamics of frames. Most of the work relied on Recurrent Neural Network (RNN) and recently attentional mechanisms have also been applied to make the model learn to focus on some frames of the video while generating each word in a describing sentence. In this paper, we focus on a sequence-to-sequence approach with temporal attention mechanism. We analyze and compare the results from different attention model configuration. By applying the temporal attention mechanism to the system, we can achieve a METEOR score of 0.310 on Microsoft Video Description dataset, which outperformed the state-of-the-art system so far.
Video event detection is a challenging problem in information and multimedia retrieval. Different from single action detection, event detection requires a richer level of semantic information from video. In order to overcome this challenge, existing solutions often represent videos using high level features such as concepts. However, concept-based representation can be confusing because it does not encode the relationship between concepts. This issue can be addressed by exploiting the co-occurrences of the concepts, however, it often leads to a very huge number of possible combinations. In this paper, we propose a new approach to obtain the relationship between concepts by exploiting the syntactic dependencies between words in the image captions. The main advantage of this approach is that it significantly reduces the number of informative combinations between concepts. We conduct extensive experiments to analyze the effectiveness of using the new dependency representation for event detection on two large-scale TRECVID Multimedia Event Detection 2013 and 2014 datasets. Experimental results show that i) Dependency features are more discriminative than concept-based features. ii) Dependency features can be combined with our current event detection system to further improve the performance. For instance, the relative improvement can be as far as 8.6% on the MEDTEST14 10Ex setting.
We designed a new annotation scheme for formalising relation structures in research papers, through the investigation of computer science papers. The annotation scheme is based on the hypothesis that identifying the role of entities and events that are described in a paper is useful for intelligent information retrieval in academic literature, and the role can be determined by the relationship between the author and the described entities or events, and relationships among them. Using the scheme, we have annotated research abstracts from the IPSJ Journal published in Japanese by the Information Processing Society of Japan. On the basis of the annotated corpus, we have developed a prototype information extraction system which has the facility to classify sentences according to the relationship between entities mentioned, to help find the role of the entity in which the searcher is interested.
We introduce the organization of the Todai Robot Project and discuss its achievements. The Todai Robot Project task focuses on benchmarking NLP systems for problem solving. This task encourages NLP-based systems to solve real high-school examinations. We describe the details of the method to manage question resources and their correct answers, answering tools and participation by researchers in the task. We also analyse the answering accuracy of the developed systems by comparing the systems answers with answers given by human test-takers.
We have created a scheme for annotating corpora designed to capture relevant aspects of factivity in verb-complement constructions. Factivity constructions are a well-known linguistic phenomenon that embed presuppositions about the state of the world into a clause. These embedded presuppositions provide implicit information about facts assumed to be true in the world, and are thus potentially valuable in areas of research such as textual entailment. We attempt to address both clear-cut cases of factivity and non-factivity, as well as account for the fluidity and ambiguous nature of some realizations of this construction. Our extensible scheme is designed to account for distinctions between claims, performatives, atypical uses of factivity, and the authority of the one making the utterance. We introduce a simple XML-based syntax for the annotation of factive verbs and clauses, in order to capture this information. We also provide an analysis of the issues which led to these annotative decisions, in the hope that these analyses will be beneficial to those dealing with factivity in a practical context.
This paper introduces our study on creating a Japanese corpus that is annotated using semantically-motivated predicate-argument structures. We propose an annotation framework based on Lexical Conceptual Structure (LCS), where semantic roles of arguments are represented through a semantic structure decomposed by several primitive predicates. As a first stage of the project, we extended Jackendoff 's LCS theory to increase generality of expression and coverage for verbs frequently appearing in the corpus, and successfully created LCS structures for 60 frequent Japanese predicates in Kyoto university Text Corpus (KTC). In this paper, we report our framework for creating the corpus and the current status of creating an LCS dictionary for Japanese predicates.
We present the UOT Machine Translation System that was used in the IWSLT-09 evaluation campaign. This year, we participated in the BTEC track for Chinese-to-English translation. Our system is based on a string-to-tree framework. To integrate deep syntactic information, we propose the use of parse trees and semantic dependencies on English sentences described respectively by Head-driven Phrase Structure Grammar and Predicate-Argument Structures. We report the results of our system on both the development and test sets.
We report the construction of a corpus for parser evaluation in the biomedical domain. A 50-abstract subset (492 sentences) of the GENIA corpus (Kim et al., 2003) is annotated with labeled head-dependent relations using the grammatical relations (GR) evaluation scheme (Carroll et al., 1998) ,which has been used for parser evaluation in the newswire domain.