COLING 2024 Papers

Well annotated corpora have been shown to have great value, both in linguistic and non-linguistic research, and in supporting machine-learning and many other non-research activities including language teaching. For minority languages, annotated corpora can help in understanding language usage norms among native and non-native speakers, providing valuable information both for lexicography and for teaching, and helping to combat the decline of speaker numbers. At the same time, minority languages suffer from having fewer available language resources than majority languages, and far less-developed annotation tooling. To date there is very little work in semantic annotation for Irish. In this paper we report on progress to date in the building of a standard tool-set for semantic annotation of Irish, including a novel method for evaluation of semantic annotation. A small corpus of Irish language data has been manually annotated with semantic tags, and manually checked. A semantic type tagging framework has then been developed using existing technologies, and using a semantic lexicon that has been built from a variety of sources. Semantic disambiguation methods have been added with a view to increasing accuracy. That framework has then been tested using the manually tagged corpus, resulting in over 90% lexical coverage and almost 80% tag accuracy. Development is ongoing as part of a larger corpus development project, and plans include expansion of the manually tagged corpus, expansion of the lexicon, and exploration of further disambiguation methods. As the first semantic tagger for Irish, to our knowledge, it is hoped that this research will form a sound basis for semantic annotation of Irish corpora in to the future.

This paper addresses the task of temporal activity localization (TAL). Although recent works have made significant progress in TAL research, almost all of them implicitly assume that the dense frame-level correspondences in each video-query pair are correctly annotated. However, in reality, such an assumption is extremely expensive and even impossible to satisfy due to subjective labeling. To alleviate this issue, in this paper, we explore a new TAL setting termed Noisy Temporal activity localization (NTAL), where a TAL model should be robust to the mixed training data with noisy moment boundaries. Inspired by the memorization effect of neural networks, we propose a novel method called Co-Teaching Regularizer (CTR) for NTAL. Specifically, we first learn a Gaussian Mixture Model to divide the mixed training data into preliminary clean and noisy subsets. Subsequently, we refine the labels of the two subsets by an adaptive prediction function so that their true positive and false positive samples could be identified. To avoid single model being prone to its mistakes learned by the mixed data, we adopt a co-teaching paradigm, which utilizes two models sharing the same framework to teach each other for robust learning. A curriculum strategy is further introduced to gradually learn the moment confidence from easy to hard. Experiments on three datasets demonstrate that our CTR is significantly more robust to the noisy training data compared to the existing methods.

Using large language models (LLMs) for machine translation via in-context learning (ICL) has become an interesting research direction of machine translation (MT) in recent years. Its main idea is to retrieve a few translation pairs as demonstrations from an additional datastore (parallel corpus) to guide translation without updating the LLMs. However, the underlying noise of retrieved demonstrations usually dramatically deteriorate the performance of LLMs. In this paper, we propose a robust method to enable LLMs to achieve robust translation with ICL. The method incorporates a multi-view approach, considering both sentence- and word-level information, to select demonstrations that effectively avoid noise. At the sentence level, a margin-based score is designed to avoid semantic noise. At the word level, word embeddings are utilized to evaluate the related tokens and change the weight of words in demonstrations. By considering both sentence- and word-level similarity, the proposed method provides fine-grained demonstrations that effectively prompt the translation of LLMs. Experimental results demonstrate the effectiveness of our method, particularly in domain adaptation.

Evidence-aware fake news detection aims to determine the veracity of a given news (i.e., claim) with external evidences. We find that existing methods lack sufficient semantic perception and are easily blinded by textual expressions. For example, they still make the same prediction after we flip the semantics of a claim, which makes them vulnerable to malicious attacks. In this paper, we propose a model-agnostic training framework to improve the semantic perception of evidence-aware fake news detection. Specifically, we first introduce two kinds of data augmentation to complement the original training set with synthetic data. The semantic-flipped augmentation synthesizes claims with similar textual expressions but opposite semantics, while the semantic-invariant augmentation synthesizes claims with the same semantics but different writing styles. Moreover, we design a novel module to learn better claim representation which is more sensitive to the semantics, and further incorporate it into a multi-objective optimization paradigm. In the experiments, we also extend the original test set of benchmark datasets with the synthetic data to better evaluate the model perception of semantics. Experimental results demonstrate that our approach significantly outperforms the state-of-the-art methods on the extended test set, while achieving competitive performance on the original one. Our source code are released at https://github.com/Xyang1998/RobustFND.

We introduce a meta dataset for few-shot relation extraction, which includes two datasets derived from existing supervised relation extraction datasets – NYT29 (Takanobu et al., 2019; Nayak and Ng, 2020) and WIKI- DATA (Sorokin and Gurevych, 2017) – as well as a few-shot form of the TACRED dataset (Sabo et al., 2021). Importantly, all these few-shot datasets were generated under realistic assumptions such as: the test relations are different from any relations a model might have seen before, limited training data, and a preponderance of candidate relation mentions that do not correspond to any of the relations of interest. Using this large resource, we conduct a comprehensive evaluation of six recent few-shot relation extraction methods, and observe that no method comes out as a clear winner. Further, the overall performance on this task is low, indicating substantial need for future research. We release all versions of the data, i.e., both supervised and few-shot, for future research.

Multi-modal sarcasm detection aims to identify whether a given sample with multi-modal information (i.e., text and image) is sarcastic, which has received increasing attention due to the rapid growth of multi-modal posts on modern social media. However, mainstream models process the input of each modality in a holistic manner, resulting in redundant and unrefined information. Moreover, the representations of different modalities are entangled in one common latent space to perform complex cross-modal interactions, neglecting the heterogeneity and distribution gap of different modalities. To address these issues, we propose a novel framework DMMD (short for Disentangled Multi-grained Multi-modal Distilling) for multi-modal sarcasm detection, which conducts multi-grained knowledge distilling (i.e., intra-subspace and inter-subspace) based on the disentangled multi-modal representations. Concretely, the representations of each modality are disentangled explicitly into modality-agnostic/specific subspaces. Then we transfer cross-modal knowledge by conducting intra-subspace knowledge distilling in a self-adaptive pattern. We also apply mutual learning to regularize the underlying inter-subspace consistency. Extensive experiments on a commonly used benchmark demonstrate the efficacy of our DMMD over cutting-edge methods. More encouragingly, visualization results indicate the multi-modal representations display meaningful distributional patterns, and we hope it will be helpful for the community of multi-modal knowledge transfer.

Large Language Models (LLMs) hold considerable promise for artificial general intelligence, given their intrinsic abilities to accomplish a wide range of open-domain tasks either independently or in tandem with specialized expert models. However, despite these capabilities, the performance of LLMs has yet to be comprehensively evaluated in realistic scenarios. To this end, in this work, we introduce a novel task, the Realistic Chinese Spell Checking (RCSC), to evaluate the effectiveness of existing methods comprehensively. In contrast to existing works that solely address Chinese character misspellings or pinyin conversions, our task aims to convert the realistic Chinese text into the corresponding correct text. The realistic Chinese text may potentially contain both Chinese misspellings and pinyin conversions. We first present the Realistic Chinese Spell Checking Benchmark (RCSCB), which consists of two subsets and contains a total of 581,657 samples. Then, we benchmark the performance of various baselines and find that all the existing methods, including instruction-based LLMs, achieve unsatisfactory results on RCSCB. To further improve the performance on RCSCB, we propose Pinyin-Enhanced Spell Checker (PESC), which is specifically designed to address pinyin-related misspellings. Experimental results demonstrate that PESC can achieve state-of-the-art performance on RCSCB. Despite the progress made, the current state-of-the-art performance is still far from satisfactory. We expect further progress on this crucial and challenging task.

Math Word Problem (MWP) is a crucial NLP task aimed at providing solutions for given mathematical descriptions. A notable sub-category of MWP is the Linear Programming Word Problem (LPWP), which holds significant relevance in real-world decision-making and operations research. While the recent rise of generative large language models (LLMs) has brought more advanced solutions to LPWPs, existing evaluation methodologies for this task still diverge from human judgment and face challenges in recognizing mathematically equivalent answers. In this paper, we introduce a novel evaluation metric rooted in graph edit distance, featuring benefits such as permutation invariance and more accurate program equivalence identification. Human evaluations empirically validate the superior efficacy of our proposed metric when particularly assessing LLM-based solutions for LPWP.

Key-value relations are prevalent in Visually-Rich Documents (VRDs), often depicted in distinct spatial regions accompanied by specific color and font styles. These non-textual cues serve as important indicators that greatly enhance human comprehension and acquisition of such relation triplets. However, current document AI approaches often fail to consider this valuable prior information related to visual and spatial features, resulting in suboptimal performance, particularly when dealing with limited examples. To address this limitation, our research focuses on few-shot relational learning, specifically targeting the extraction of key-value relation triplets in VRDs. Given the absence of a suitable dataset for this task, we introduce two new few-shot benchmarks built upon existing supervised benchmark datasets. Furthermore, we propose a variational approach that incorporates relational 2D-spatial priors and prototypical rectification techniques. This approach aims to generate relation representations that are more aware of the spatial context and unseen relation in a manner similar to human perception. Experimental results demonstrate the effectiveness of our proposed method by showcasing its ability to outperform existing methods. This study also opens up new possibilities for practical applications.

In textual question answering (TQA) systems, complex questions often require retrieving multiple textual fact chains with multiple reasoning steps. While existing benchmarks are limited to single-chain or single-hop retrieval scenarios. In this paper, we propose to conduct Graph-Hop —— a novel multi-chains and multi-hops retrieval and reasoning paradigm in complex question answering. We construct a new benchmark called ReasonGraphQA, which provides explicit and fine-grained evidence graphs for complex question to support comprehensive and detailed reasoning. In order to further study how graph-based evidential reasoning can be performed, we explore what form of Graph-Hop works best for generating textual evidence explanations in knowledge reasoning and question answering. We have thoroughly evaluated existing evidence retrieval and reasoning models on the ReasonGraphQA. Experiments highlight Graph-Hop is a promising direction for answering complex questions, but it still has certain limitations. We have further studied mitigation strategies to meet these challenges and discuss future directions.

Large language models (LLMs) have achieved significant performance in various natural language reasoning tasks. However, they still struggle with performing first-order logic reasoning over formal logical theories expressed in natural language. This is because the previous LLMs-based reasoning systems have the theoretical incompleteness issue. As a result, it can only address a limited set of simple reasoning problems, which significantly decreases their generalization ability. To address this issue, we propose a novel framework, named Generalizable and Faithful Reasoner (GFaiR), which introduces the paradigm of resolution refutation. Resolution refutation has the capability to solve all first-order logic reasoning problems by extending reasoning rules and employing the principle of proof by contradiction, so our system’s completeness can be improved by introducing resolution refutation. Experimental results demonstrate that our system outperforms previous works by achieving state-of-the-art performances in complex scenarios while maintaining performances in simple scenarios. Besides, we observe that GFaiR is faithful to its reasoning process.

Recent advances of incorporating layout information, typically bounding box coordinates, into pre-trained language models have achieved significant performance in entity recognition from document images. Using coordinates can easily model the position of each token, but they are sensitive to manipulations in document images (e.g., shifting, rotation or scaling) which are common in real scenarios. Such limitation becomes even worse when the training data is limited in few-shot settings. In this paper, we propose a novel framework, LAGER, which leverages the topological adjacency relationship among the tokens through learning their relative layout information with graph neural networks. Specifically, we consider the tokens in the documents as nodes and formulate the edges based on the topological heuristics. Such adjacency graphs are invariant to affine transformations, making it robust to the common image manipulations. We incorporate these graphs into the pre-trained language model by adding graph neural network layers on top of the language model embeddings. Extensive experiments on two benchmark datasets show that LAGER significantly outperforms strong baselines under different few-shot settings and also demonstrate better robustness to manipulations.

The assessment of explainability in Legal Judgement Prediction (LJP) systems is of paramount importance in building trustworthy and transparent systems, particularly considering the reliance of these systems on factors that may lack legal relevance or involve sensitive attributes. This study delves into the realm of explainability and fairness in LJP models, utilizing Swiss Judgement Prediction (SJP), the only available multilingual LJP dataset. We curate a comprehensive collection of rationales that ‘support’ and ‘oppose’ judgement from legal experts for 108 cases in German, French, and Italian. By employing an occlusion-based explainability approach, we evaluate the explainability performance of state-of-the-art monolingual and multilingual BERT-based LJP models, as well as models developed with techniques such as data augmentation and cross-lingual transfer, which demonstrated prediction performance improvement. Notably, our findings reveal that improved prediction performance does not necessarily correspond to enhanced explainability performance, underscoring the significance of evaluating models from an explainability perspective. Additionally, we introduce a novel evaluation framework, Lower Court Insertion (LCI), which allows us to quantify the influence of lower court information on model predictions, exposing current models’ biases.

Despite being widely spoken, dialectal variants of languages are frequently considered low in resources due to lack of writing standards and orthographic inconsistencies. As a result, training natural language understanding (NLU) systems relies primarily on standard language resources leading to biased and inequitable NLU technology that underserves dialectal speakers. In this paper, we propose to address this problem through a framework composed of a dialect identification model that is used to obtain targeted training data augmentation for under-represented dialects, in an effort to debias NLU model for dialectal cohorts in NLU systems. We conduct experiments on two dialect rich non-English languages: Arabic and German, using large-scale commercial NLU datasets as well as open-source datasets. Results show that such framework can provide insights on dialect disparity in real-world NLU systems and targeted data argumentation can help narrow the model’s performance gap between standard language speakers and dialect speakers.

Similar to humans, animals make extensive use of verbal and non-verbal forms of communication, including a large range of audio signals. In this paper, we address dog vocalizations and explore the use of self-supervised speech representation models pre-trained on human speech to address dog bark classification tasks that find parallels in human-centered tasks in speech recognition. We specifically address four tasks: dog recognition, breed identification, gender classification, and context grounding. We show that using speech embedding representations significantly improves over simpler classification baselines. Further, we also find that models pre-trained on large human speech acoustics can provide additional performance boosts on several tasks.

Conversations exhibit significant variation when different styles are employed by participants, often leading to subpar performance when a dialogue model is exclusively trained on single-style datasets. We present a cost-effective methodology for generating multi-style conversations, which can be used in the development of conversational agents. This methodology only assumes the availability of a conversational domain, such as a knowledge base, and leverages the generative capabilities of large language models. In a pilot study focused on the generation aspect of task-oriented dialogues, we extended the well-known MultiWOZ dataset to encompass multi-style variations. Our findings highlight two key experimental outcomes: (i) these novel resources pose challenges for current single-style models, and (ii) multi-style resources enhance the dialogue model’s resilience to stylistic variations.

Contemporary NLP has rapidly progressed from feature-based classification to fine-tuning and prompt-based techniques leveraging large language models. Many of these techniques remain understudied in the context of real-world, clinically enriched spontaneous dialogue. We fill this gap by systematically testing the efficacy and overall performance of a wide variety of NLP techniques ranging from feature-based to in-context learning on transcribed speech collected from patients with bipolar disorder, schizophrenia, and healthy controls taking a focused, clinically-validated language test. We observe impressive utility of a range of feature-based and language modeling techniques, finding that these approaches may provide a plethora of information capable of upholding clinical truths about these subjects. Building upon this, we establish pathways for future research directions in automated detection and understanding of psychiatric conditions.

Readability is a crucial characteristic of texts, greatly influencing comprehension and reading efficacy. Unfortunately, limited research is available for less-resourced languages, especially for young populations where its impact is even higher. This paper introduces a new readability tool for children’s literature in the Romanian language, explicitly targeting primary school students aged 7-11. The tool consists of a digital repository of school reading texts (self-compiled corpus) and a text analysis interface that generates automatic readability reports for uploaded short texts. The methodology involves extracting, testing, and calibrating a readability formula for Romanian using the children’s literature corpus. Related work on readability and readability tools is discussed, followed by a description of the children’s literature corpus and the platform functionalities. The first steps are presented towards validating the readability formula for children’s literature in Romanian using the ReaderBench framework, while calibration variables relevant to the Romanian language and children’s literature are examined. Currently, no existing platform integrates a research-based readability formula for the Romanian language, making this tool unique. Overall, this research contributes to applied corpus linguistics and Digital Humanities studies and offers a valuable resource for educators, parents, and children in accessing age-appropriate and readable texts.

In recent research, significant advancements have been achieved in tool learning for large language models. Looking towards future advanced studies, the issue of fully autonomous tool utilization is particularly intriguing: given only a query, language models can autonomously decide whether to employ a tool, which specific tool to select, and how to utilize these tools, all without needing any tool-specific prompts within the context. To achieve this, we introduce a unified, efficient, and scalable framework for fine-tuning language models. Based on the degree of tool dependency, we initially categorize queries into three distinct types. By transforming the entire process into a sequential decision-making problem through conditional probability decomposition, our approach unifies the three types and autoregressively generates decision processes. Concurrently, we’ve introduced an “instruct, execute, and reformat” strategy specifically designed for efficient data annotation. Through end-to-end training on the annotated dataset comprising 26 diverse APIs, the model demonstrates a level of self-awareness, automatically seeking tool assistance when necessary. It significantly surpasses original instruction-tuned open-source language models and GPT-3.5/4 on multiple evaluation metrics. To address real-world scalability needs, we’ve enhanced our framework with a dynamic rehearsal strategy for continual learning, proven to require minimal new annotations to exhibit remarkable performance.