COLING 2024 Papers

Image-to-text generation involves automatically generating descriptive text from images and has applications in medical report generation. However, traditional approaches often exhibit a semantic gap between visual and textual information. In this paper, we propose a multi-task learning framework to leverage both visual and non-imaging data for generating radiology reports. Along with chest X-ray images, 10 additional features comprising numeric, binary, categorical, and text data were incorporated to create a unified representation. The model was trained to generate text, predict the degree of patient severity, and identify medical findings. Multi-task learning, especially with text generation prioritisation, improved performance over single-task baselines across language generation metrics. The framework also mitigated overfitting in auxiliary tasks compared to single-task models. Qualitative analysis showed logically coherent narratives and accurate identification of findings, though some repetition and disjointed phrasing remained. This work demonstrates the benefits of multi-modal, multi-task learning for image-to-text generation applications.

Pretrained language models excel in various natural language processing tasks but often neglect the integration of different scripts within a language, constraining their ability to capture richer semantic information, such as in Hindi. In this work, we present a dual-script enhanced feature representation method for Hindi. We combine single-script features from Devanagari and Romanized Hindi Roberta using concatenation, addition, cross-attention, and convolutional networks. The experiment results show that using a dual-script approach significantly improves model performance across various tasks. The addition fusion technique excels in sequence generation tasks, while for text classification, the CNN-based dual-script enhanced representation performs best with longer sentences, and the addition fusion technique is more effective for shorter sequences. Our approach shows significant advantages in multiple natural language processing tasks, providing a new perspective on feature representation for Hindi. Our code has been released on https://github.com/JohnnyChanV/Hindi-Fusion.

As pretrained language model emerge and consistently develop, prompt-based training has become a well-studied paradigm to improve the exploitation of models for many natural language processing tasks. Furthermore, prompting demonstrates great performance compared to conventional fine-tuning in scenarios with limited annotated data, such as zero-shot or few-shot situations. Verbalizers are crucial in this context, as they help interpret masked word distributions generated by language models into output predictions. This study introduces a benchmarking approach to assess three common baselines of verbalizers for topic classification in few-shot learning scenarios. Additionally, we find that increasing the number of label words for automatic label word searching enhances model performance. Moreover, we investigate the effectiveness of template assembling with various aggregation strategies to develop stronger classifiers that outperform models trained with individual templates. Our approach achieves comparable results to prior research while using significantly fewer resources. Our code is available at https://github.com/quang-anh-nguyen/verbalizer_benchmark.git.

Predicting emotions elicited by news headlines can be challenging as the task is largely influenced by the varying nature of people’s interpretations and backgrounds. Previous works have explored classifying discrete emotions directly from news headlines. We provide a different approach to tackling this problem by utilizing people’s explanations of their emotion, written in free-text, on how they feel after reading a news headline. Using the dataset BU-NEmo+ (Gao et al., 2022), we found that for emotion classification, the free-text explanations have a strong correlation with the dominant emotion elicited by the headlines. The free-text explanations also contain more sentimental context than the news headlines alone and can serve as a better input to emotion classification models. Therefore, in this work we explored generating emotion explanations from headlines by training a sequence-to-sequence transformer model and by using pretrained large language model, ChatGPT (GPT-4). We then used the generated emotion explanations for emotion classification. In addition, we also experimented with training the pretrained T5 model for the intermediate task of explanation generation before fine-tuning it for emotion classification. Using McNemar’s significance test, methods that incorporate GPT-generated free-text emotion explanations demonstrated significant improvement (P-value < 0.05) in emotion classification from headlines, compared to methods that only use headlines. This underscores the value of using intermediate free-text explanations for emotion prediction tasks with headlines.

Efforts to leverage deep learning models in low-resource regimes have led to numerous augmentation studies. However, the direct application of methods, such as mixup and cutout, is limited due to the discrete characteristics of the textual data. While methods using pre trained language models have exhibited good efficiency, they require additional considerations for robustness. Inspired by recent studies on decision boundaries, this paper proposes a decision-boundary-aware data augmentation strategy to enhance robustness using pretrained language models. The proposed technique first focuses on shifting the latent features closer to the decision boundary, followed by reconstruction to generate an ambiguous version with a soft label. Additionally, mid-K sampling is suggested to enhance the diversity of the generated sentences. This paper demonstrates the performance of the proposed augmentation strategy compared to other methods through extensive experiments. Furthermore, the ablation study demonstrates the effect of soft labels and mid-K sampling and the extensibility of the method with curriculum data augmentation.

In low-resource Named Entity Recognition (NER) scenarios, only a limited quantity of strongly labeled data is available, while a vast amount of weakly labeled data can be easily acquired through distant supervision. However, weakly labeled data may fail to improve the model performance or even harm it due to the inevitable noise. While training on noisy data, only certain parameters are essential for model learning, termed safe parameters, whereas the other parameters tend to fit noise. In this paper, we propose a noise-robust learning framework where safe parameters can be identified with guidance from the small set of strongly labeled data, and non-safe parameters are suppressed during training on weakly labeled data for better generalization. Our method can effectively mitigate the impact of noise in weakly labeled data, and it can be easily integrated with data level noise-robust learning methods for NER. We conduct extensive experiments on multiple datasets and the results show that our approach outperforms the state-of-the-art methods.

Existing cross-document event coreference resolution models, which either compute mention similarity directly or enhance mention representation by extracting event arguments (such as location, time, agent, and patient), lackingmthe ability to utilize document-level information. As a result, they struggle to capture long-distance dependencies. This shortcoming leads to their underwhelming performance in determining coreference for the events where their argument information relies on long-distance dependencies. In light of these limitations, we propose the construction of document-level Rhetorical Structure Theory (RST) trees and cross-document Lexical Chains to model the structural and semantic information of documents. Subsequently, cross-document heterogeneous graphs are constructed and GAT is utilized to learn the representations of events. Finally, a pair scorer calculates the similarity between each pair of events and co-referred events can be recognized using standard clustering algorithm. Additionally, as the existing cross-document event coreference datasets are limited to English, we have developed a large-scale Chinese cross-document event coreference dataset to fill this gap, which comprises 53,066 event mentions and 4,476 clusters. After applying our model on the English and Chinese datasets respectively, it outperforms all baselines by large margins.

Court View Generation (CVG) is a challenging task in the field of Legal Artificial Intelligence (LegalAI), which aims to generate court views based on the plaintiff claims and the fact descriptions. While Pretrained Language Models (PLMs) have showcased their prowess in natural language generation, their application to the complex, knowledge-intensive domain of CVG often reveals inherent limitations. In this paper, we present a novel approach, named Knowledge Injection and Guidance (KIG), designed to bolster CVG using PLMs. To efficiently incorporate domain knowledge during the training stage, we introduce a knowledge-injected prompt encoder for prompt tuning, thereby reducing computational overhead. Moreover, to further enhance the model’s ability to utilize domain knowledge, we employ a generating navigator, which dynamically guides the text generation process in the inference stage without altering the model’s architecture, making it readily transferable. Comprehensive experiments on real-world data demonstrate the effectiveness of our approach compared to several established baselines, especially in the responsivity of claims, where it outperforms the best baseline by 11.87%.

Large Language Models (LLMs) have recently made significant advances in code generation through the ‘Chain-of-Thought’ prompting technique. This technique empowers the model to autonomously devise “solution plans” to tackle intricate programming challenges, thereby improving its performance in code generation. Nevertheless, smaller models have been struggling to keep up with LLMs in deducing these plans, adversely affecting their code generation capabilities. Given the considerable size and associated deployment costs, along with concerns about data security, many teams opt for deploying smaller models for code generation. Consequently, there arises a compelling need for transferring LLMs’ code generation reasoning abilities to the smaller models. In this paper, we propose the CodePLAN framework, which aims to transfer LLMs’ reasoning capabilities to smaller models through distillation. We adopt a multi-task learning approach, jointly undertaking code generation and solution plan generation tasks, to enhance the code generation capabilities of smaller model. To ensure the superior quality of the solution plans, we advocate for the utilization of backward reasoning and plan sampling strategies. Our experiments show that in comparison to the conventional fine-tuning approach, our approach improves the smaller model’s code generation performance (measured in pass@1 metric) by over 130% on the challenging APPS benchmark.

In information retrieval, facet identification of a user query is an important task. If a search service can recognize the facets of a user’s query, it has the potential to offer users a much broader range of search results. Previous studies can enhance facet prediction by leveraging retrieved documents and related queries obtained through a search engine. However, there are challenges in extending it to other applications when a search engine operates as part of the model. First, search engines are constantly updated. Therefore, additional information may change during training and test, which may reduce performance. The second challenge is that public search engines cannot search for internal documents. Therefore, a separate search system needs to be built to incorporate documents from private domains within the company. We propose two strategies that focus on a framework that can predict facets by taking only queries as input without a search engine. The first strategy is multi-task learning to predict SERP. By leveraging SERP as a target instead of a source, the proposed model deeply understands queries without relying on external modules. The second strategy is to enhance the facets by combining Large Language Model (LLM) and the small model. Overall performance improves when small model and LLM are combined rather than facet generation individually.

Aspect-category-based sentiment analysis (ACSA), which aims to identify aspect categories and predict their sentiments has been intensively studied due to its wide range of NLP applications. Most approaches mainly utilize intrasentential features. However, a review often includes multiple different aspect categories, and some of them do not explicitly appear in the review. Even in a sentence, there is more than one aspect category with its sentiments, and they are entangled intra-sentence, which makes the model fail to discriminately preserve all sentiment characteristics. In this paper, we propose an enhanced coherence-aware network with hierarchical disentanglement (ECAN) for ACSA tasks. Specifically, we explore coherence modeling to capture the contexts across the whole review and to help the implicit aspect and sentiment identification. To address the issue of multiple aspect categories and sentiment entanglement, we propose a hierarchical disentanglement module to extract distinct categories and sentiment features. Extensive experimental and visualization results show that our ECAN effectively decouples multiple categories and sentiments entangled in the coherence representations and achieves state-of-the-art (SOTA) performance. Our codes and data are available online: https://github.com/cuijin-23/ECAN.

The widespread use of pre-trained language models (PLMs) in natural language processing (NLP) has greatly improved performance outcomes. However, these models’ vulnerability to adversarial attacks (e.g., camouflaged hints from drug dealers), particularly in the Chinese language with its rich character diversity/variation and complex structures, hatches vital apprehension. In this study, we propose a novel method, CHinese vAriatioN Graph Enhancement (CHANGE), to increase the robustness of PLMs against character variation attacks in Chinese content. CHANGE presents a novel approach to incorporate a Chinese character variation graph into the PLMs. Through designing different supplementary tasks utilizing the graph structure, CHANGE essentially enhances PLMs’ interpretation of adversarially manipulated text. Experiments conducted in a multitude of NLP tasks show that CHANGE outperforms current language models in combating against adversarial attacks and serves as a valuable contribution to robust language model research. Moreover, these findings highlight the substantial potential of graph-guided pre-training strategies for real-world applications.

Natural languages show a tendency to minimize the linear distance between heads and their dependents in a sentence, known as dependency length minimization (DLM). Such a preference, however, has not been consistently replicated with neural agent simulations. Comparing the behavior of models with that of human learners can reveal which aspects affect the emergence of this phenomenon. In this work, we investigate the minimal conditions that may lead neural learners to develop a DLM preference. We add three factors to the standard neural-agent language learning and communication framework to make the simulation more realistic, namely: (i) the presence of noise during listening, (ii) context-sensitivity of word use through non-uniform conditional word distributions, and (iii) incremental sentence processing, or the extent to which an utterance’s meaning can be guessed before hearing it entirely. While no preference appears in production, we show that the proposed factors can contribute to a small but significant learning advantage of DLM for listeners of verb-initial languages.

We focus on the problem of parsing procedural text into fine-grained flow graphs that encode actions and entities, as well as their interactions. Specifically, we focus on parsing cooking recipes, and address a few limitations of existing parsers. Unlike SOTA approaches to flow graph parsing that work in two separate stages identifying actions and entities (tagging) and encoding their interactions via connecting edges (graph generation). we propose an end-to-end multi-task framework that simultaneously performs tagging and graph generation. In addition, due to the end-to-end nature of our proposed model, we can unify the input representation, and moreover can use compact encoders, resulting in small models with significantly fewer parameters than SOTA models. Another key challenge in training flow graph parsers is the lack of sufficient annotated data, due to the costly nature of the fine-grained annotations. We address this problem by taking advantage of the abundant unlabelled recipes, and show that pre-training on automatically-generated noisy silver annotations (from unlabelled recipes) results in a large improvement in flow graph parsing.

Abstract Meaning Representation (AMR) is a general-purpose meaning representation that has become popular for its clear structure, ease of annotation and available corpora, and overall expressiveness. While AMR was designed to represent sentence meaning in English text, recent research has explored its adaptation to broader domains, including documents, dialogues, spatial information, cross-lingual tasks, and gesture. In this paper, we present an annotated corpus of multimodal (speech and gesture) AMR in a task-based setting. Our corpus is multilayered, containing temporal alignments to both the speech signal and to descriptions of gesture morphology. We also capture coreference relationships across modalities, enabling fine-grained analysis of how the semantics of gesture and natural language interact. We discuss challenges that arise when identifying cross-modal coreference and anaphora, as well as in creating and evaluating multimodal corpora in general. Although we find AMR’s abstraction away from surface form (in both language and gesture) occasionally too coarse-grained to capture certain cross-modal interactions, we believe its flexibility allows for future work to fill in these gaps. Our corpus and annotation guidelines are available at https://github.com/klai12/encoding-gesture-multimodal-dialogue.

To provide effective support, it is essential for a skilled supporter to emotionally resonate with the help-seeker’s current emotional state. In conversational interactions, this emotional alignment is further influenced by the comforting strategies employed by the supporter. Different strategies guide the interlocutors to align their emotions in nuanced patterns. However, the incorporation of strategy into emotional alignment in the context of emotional support agents remains underexplored. To address this limitation, we propose an improved emotional support agent called Emstremo. Emstremo aims to achieve strategic control of emotional alignment by perceiving and responding to the user’s emotions. Our system’s state-of-the-art performance emphasizes the importance of integrating emotions and strategies in modeling conversations that provide emotional support.

The study delves into the construction of entailment trees for science question answering (SQA), employing a novel framework termed Tree-structured Entailment Reasoning (TER). Current research on entailment tree construction presents significant challenges, primarily due to the ambiguities and similarities among candidate science facts, which considerably complicate the fact retrieval process. Moreover, the existing models exhibit limitations in effectively modeling the sequence of reasoning states, understanding the intricate relations between neighboring entailment tree nodes, and generating intermediate conclusions. To this end, we explore enhancing the TER performance from three aspects: First, improving retrieval capabilities by modeling and referring to the chained reasoning states; Second, enhancing TER by infusing knowledge that bridges the gap between reasoning types and rhetorical relations. Third, exploring a task-specific large language model tuning scheme to mitigate deficiencies in intermediate conclusion generation. Experiments on the English EntailmentBank demonstrate the effectiveness of the proposed methods in augmenting the quality of tree-structured entailment reasoning to a certain extent.

Knowledge-intensive tasks pose a significant challenge for Machine Learning (ML) techniques. Commonly adopted methods, such as Large Language Models (LLMs), often exhibit limitations when applied to such tasks. Nevertheless, there have been notable endeavours to mitigate these challenges, with a significant emphasis on augmenting LLMs through Knowledge Graphs (KGs). While KGs provide many advantages for representing knowledge, their development costs can deter extensive research and applications. Addressing this limitation, we introduce a framework for enriching embeddings of small-scale domain-specific Knowledge Graphs with well-established general-purpose KGs. Adopting our method, a modest domain-specific KG can benefit from a performance boost in downstream tasks when linked to a substantial general-purpose KG. Experimental evaluations demonstrate a notable enhancement, with up to a 44% increase observed in the Hits@10 metric. This relatively unexplored research direction can catalyze more frequent incorporation of KGs in knowledge-intensive tasks, resulting in more robust, reliable ML implementations, which hallucinates less than prevalent LLM solutions.

In this paper, we present the development of a digital Oneida verb conjugator through using the Gramble framework. This project is a collaborative effort with the Twatati Adult Oneida Language program. Oneida is a polysynthetic North American Indigenous language. Its verb roots can be conjugated with multiple affixes, and long verbal complexes can be used as utterances. Each Oneida affix encodes important grammatical information, and its form often varies based on various factors, such as its position in the utterance and its phonological environment. The distinct morphosyntactic structures complicate acquisition of the language by learners who are native speakers of English. With an alarmingly small number of native speakers of Oneida, supporting and accelerating adult second language leaners’ acquisition process has become a pressing necessity. The Oneida verb conjugator can demonstrate its users the correct conjugations of verbs and can also let learners generate practice materials tailored to their unique learning trajectories. This paper presents the preliminary stages and outcomes of the project and outlines the areas for improvement to be addressed in our subsequent endeavors.