COLING 2024 论文列表

Event relation extraction (ERE) is a critical and fundamental challenge for natural language processing. Existing work mainly focuses on directly modeling the entire document, which cannot effectively handle long-range dependencies and information redundancy. To address these issues, we propose a cluster-aware compression method for improving event relation extraction (TacoERE), which explores a compression-then-extraction paradigm. Specifically, we first introduce document clustering for modeling event dependencies. It splits the document into intra- and inter-clusters, where intra-clusters aim to enhance the relations within the same cluster, while inter-clusters attempt to model the related events at arbitrary distances. Secondly, we utilize cluster summarization to simplify and highlight important text content of clusters for mitigating information redundancy and event distance. We have conducted extensive experiments on both pre-trained language models, such as RoBERTa, and large language models, such as ChatGPT and GPT-4, on three ERE datasets, i.e., MAVEN-ERE, EventStoryLine and HiEve. Experimental results demonstrate that TacoERE is an effective method for ERE.

Code search with natural language helps us reuse existing code snippets. Thanks to the Transformer-based pretraining models, the performance of code search has been improved significantly. However, due to the quadratic complexity of multi-head self-attention, there is a limit on the input token length. For efficient training on standard GPUs like V100, existing pretrained code models, including GraphCodeBERT, CodeBERT, RoBERTa (code), take the first 256 tokens by default, which makes them unable to represent the complete information of long code that is greater than 256 tokens. To tackle the long code problem, we propose a new baseline SEA (Split, Encode and Aggregate), which splits long code into code blocks, encodes these blocks into embeddings, and aggregates them to obtain a comprehensive long code representation. With SEA, we could directly use Transformer-based pretraining models to model long code without changing their internal structure and re-pretraining. We also compare SEA with sparse Trasnformer methods. With GraphCodeBERT as the encoder, SEA achieves an overall mean reciprocal ranking score of 0.785, which is 10.1% higher than GraphCodeBERT on the CodeSearchNet benchmark, justifying SEA as a strong baseline for long code search.

KEPLMs are pre-trained models that utilize external knowledge to enhance language understanding. Previous language models facilitated knowledge acquisition by incorporating knowledge-related pre-training tasks learned from relation triples in knowledge graphs. However, these models do not prioritize learning embeddings for entity-related tokens. Updating all parameters in KEPLM is computationally demanding. This paper introduces TRELM, a Robust and Efficient Pre-training framework for Knowledge-Enhanced Language Models. We observe that text corpora contain entities that follow a long-tail distribution, where some are suboptimally optimized and hinder the pre-training process. To tackle this, we employ a robust approach to inject knowledge triples and employ a knowledge-augmented memory bank to capture valuable information. Moreover, updating a small subset of neurons in the feed-forward networks (FFNs) that store factual knowledge is both sufficient and efficient. Specifically, we utilize dynamic knowledge routing to identify knowledge paths in FFNs and selectively update parameters during pre-training. Experimental results show that TRELM achieves at least a 50% reduction in pre-training time and outperforms other KEPLMs in knowledge probing tasks and multiple knowledge-aware language understanding tasks.

In this paper, we introduce an innovative pre-training framework TP-Link, which aims to improve context-dependent Text-to-SQL Parsing by leveraging Linking information. This enhancement is achieved through better representation of both natural language utterances and the database schema, ultimately facilitating more effective text-to-SQL conversations. We present two novel pre-training objectives: (i) utterance linking prediction (ULP) task that models intricate syntactic relationships among natural language utterances in context-dependent text-to-SQL scenarios, and (ii) schema linking prediction (SLP) task that focuses on capturing fine-grained schema linking relationships between the utterances and the database schema. Extensive experiments demonstrate that our proposed TP-Link achieves state-of-the-art performance on two leading downstream benchmarks (i.e., SParC and CoSQL).

End-to-end multimodal aspect-based sentiment analysis (MABSA) combines multimodal aspect terms extraction (MATE) with multimodal aspect sentiment classification (MASC), aiming to simultaneously extract aspect words and classify the sentiment polarity of each aspect. However, existing MABSA methods have overlooked two issues: (i) They only focus on fusing image regional information and textual words for two subtasks of MABSA. Whereas, MATE subtask relies more on global image information to assist in obtaining the quantity and attributes of aspects. Ignoring the integration with global information may affect the performance of MABSA methods. (ii) They fail to take advantage of target information. Nevertheless, the fine-grained details of targets are important for classifying sentiments of aspects. To solve these problems, we propose a Target-oriented Multi-grained Fusion Network(TMFN). It fuses text information with global coarse-grained image information for MATE subtask and with fine-grained image information for MASC subtask. In addition, a target-oriented feature alignment (TOFA) module is designed to enhance target-related information in image features with target details. In such a way, image features will contain more target emotional-related information which is beneficial to sentiment classification. Extensive experiments show that our method outperforms state-of-the-art methods on two benchmark datasets.

The absence of explicitly tailored, accessible annotated datasets for educational purposes presents a notable obstacle for NLP tasks in languages with limited resources. This study initially explores the feasibility of using machine translation (MT) to convert an existing dataset into a Tigrinya dataset in SQuAD format. As a result, we present TIGQA, an expert-annotated dataset containing 2,685 question-answer pairs covering 122 diverse topics such as climate, water, and traffic. These pairs are from 537 context paragraphs in publicly accessible Tigrinya and Biology books. Through comprehensive analyses, we demonstrate that the TIGQA dataset requires skills beyond simple word matching, requiring both single-sentence and multiple-sentence inference abilities. We conduct experiments using state-of-the-art MRC methods, marking the first exploration of such models on TIGQA. Additionally, we estimate human performance on the dataset and juxtapose it with the results obtained from pre-trained models. The notable disparities between human performance and the best model performance underscore the potential for fu- ture enhancements to TIGQA through continued research. Our dataset is freely accessible via the provided link to encourage the research community to address the challenges in the Tigrinya MRC. Keywords: Tigrinya QA dataset, Low resource QA dataset, domain specific QA

Responding with multimodal content has been recognized as one of the essential functionalities of intelligent conversational agents. However, existing research on multimodal dialogues primarily focuses on two topics: (1) textual response generation that ground the conversation on a given image; and (2) visual response selection based on the dialogue context. In light of the aforementioned gap, we propose mulTImodal GEnerator for dialogue Response (TIGER), a unified generative model framework for multimodal dialogue response generation. Through extensive experiments, TIGER has demonstrated new state-of-the-art results, providing users with an enhanced conversational experience. A multimodal dialogue system based on TIGER is available at https://github.com/friedrichor/TIGER. A video demonstrating the system is available at https://www.youtube.com/watch?v=Kd0CMwDs8Rk.

Speech entity linking amis to recognize mentions from speech and link them to entities in knowledge bases. Previous work on entity linking mainly focuses on visual context and text context. In contrast, speech entity linking focuses on audio context. In this paper, we first propose the speech entity linking task. To facilitate the study of this task, we propose the first speech entity linking dataset, TED-EL. Our corpus is a high-quality, human-annotated, audio, text, and mention-entity pair parallel dataset derived from Technology, Entertainment, Design (TED) talks and includes a wide range of entity types (24 types). Based on TED-EL, we designed two types of models: ranking-based and generative speech entity linking models. We conducted experiments on the TED-EL dataset for both types of models. The results show that the ranking-based models outperform the generative models, achieving an F1 score of 60.68%.

Few-shot nested named entity recognition (NER), identifying named entities that are nested with a small number of labeled data, has attracted much attention. Recently, a span-based method based on three stages ( focusing, bridging and prompting) has been proposed for few-shot nested NER. However, such a span-based approach for few-shot nested NER suffers from two challenges: 1) error propagation because of its 3-stage pipeline-based framework; 2) ignoring the relationship between inner and outer entities, which is crucial for few-shot nested NER. Therefore, in this work, we propose a two-stage approach with a controllable attention soft prompt for few-shot nested named entity recognition (TECA). It consists of two components: span part identification and entity mention recognition. The span part identification provides possible entity mentions without an extra filtering module. The entity mention recognition pays fine-grained attention to the inner and outer entities and the corresponding adjacent context through the controllable attention soft prompt to classify the candidate entity mentions. Experimental results show that the TECA approach achieves state-of-the-art performance consistently on the four benchmark datasets (ACE2004, ACE2005, GENIA, and KBP2017) and outperforms several competing baseline models on F1-score by 5.62% on ACE04, 5.11% on ACE05, 3.41% on KBP2017 and 0.7% on GENIA on the 10-shot setting.

Knowledge Distillation (KD) serves as an efficient method for transferring language knowledge from open-source large language models (LLMs) to more computationally efficient models. However, challenges arise when attempting to apply vanilla KD methods to transfer knowledge from closed-source Multilingual Neural Machine Translation (MNMT) models based on LLMs. In this scenario, the soft labels and training data are not accessible, making it difficult to achieve effective knowledge transfer. To address this issue, this paper proposes a Teacher Assistant enhanced Knowledge Distillation (TAeKD) method to augment the knowledge transfer capacity from closed-source MNMT models. Specifically, TAeKD designs a fusion model that integrates translation outputs from multiple closed-source models to generate soft labels and training samples. Furthermore, a quality assessment learning mechanism is introduced to enhance the generalization of the fusion model and elevate the quality of the fusion data used to train the student model. To facilitate research on knowledge transfer from MNMT models, we also introduce FuseData, a benchmark consisting of a blend of translations from multiple closed-source systems. The experimental results show that TAeKD outperforms the previous state-of-the-art KD methods on both WMT22 and FLORES-101 test sets.

As a cross-modal task, visual storytelling aims to generate a story for an ordered image sequence automatically. Different from the image captioning task, visual storytelling requires not only modeling the relationships between objects in the image but also mining the connections between adjacent images. Recent approaches primarily utilize either end-to-end frameworks or multi-stage frameworks to generate relevant stories, but they usually overlook latent topic information. In this paper, in order to generate a more coherent and relevant story, we propose a novel method, Topic Aware Reinforcement Network for VIsual StoryTelling (TARN-VIST). In particular, we pre-extracted the topic information of stories from both visual and linguistic perspectives. Then we apply two topic-consistent reinforcement learning rewards to identify the discrepancy between the generated story and the human-labeled story so as to refine the whole generation process. Extensive experimental results on the VIST dataset and human evaluation demonstrate that our proposed model outperforms most of the competitive models across multiple evaluation metrics.

In recent years, there has been a significant increase in interest in developing Spoken Language Understanding (SLU) systems. SLU involves extracting a list of semantic information from the speech signal. A major issue for SLU systems is the lack of sufficient amount of bi-modal (audio and textual semantic annotation) training data. Existing SLU resources are mainly available in high-resource languages such as English, Mandarin and French. However, one of the current challenges concerning low-resourced languages is data collection and annotation. In this work, we present a new freely available corpus, named TARIC-SLU, composed of railway transport conversations in Tunisian dialect that is continuously annotated in dialogue acts and slots. We describe the semantic model of the dataset, the data and experiments conducted to build ASR-based and SLU-based baseline models. To facilitate its use, a complete recipe, including data preparation, training and evaluation scripts, has been built and will be integrated to SpeechBrain, a popular open-source conversational AI toolkit based on PyTorch.

Processing tabular data holds significant importance across various domains and applications. This study investigates the performance and limitations of fine-tuned models for tabular data analysis, specifically focusing on using fine-tuning mechanics on an English model towards a potential German model. The validation of the effectiveness of the transfer learning approach compares the performance of the fine-tuned German model and of the original English model on test data from the German training set. A potential shortcut that translates the German test data into English serves for comparison. Results reveal that the fine-tuned model outperforms the original model significantly, demonstrating the effectiveness of transfer learning even for a limited amount of training data. One also observes that the English model can effectively process translated German tabular data, albeit with a slight accuracy drop compared to fine-tuning. The model evaluation extends to real-world data extracted from the sustainability reports of a financial institution. The fine-tuned model proves superior in extracting knowledge from these training-unrelated tables, indicating its potential applicability in practical scenarios. This paper also releases the first manually annotated dataset for German Table Question Answering and the related annotation tool.

Twitter has emerged as a global hub for engaging in online conversations and as a research corpus for various disciplines that have recognized the significance of its user-generated content. Argument mining is an important analytical task for processing and understanding online discourse. Specifically, it aims to identify the structural elements of arguments, denoted as information and inference. These elements, however, are not static and may require context within the conversation they are in, yet there is a lack of data and annotation frameworks addressing this dynamic aspect on Twitter. We contribute TACO, the first dataset of Twitter Arguments utilizing 1,814 tweets covering 200 entire COnversations spanning six heterogeneous topics annotated with an agreement of 0.718 Krippendorff’s α among six experts. Second, we provide our annotation framework, incorporating definitions from the Cambridge Dictionary, to define and identify argument components on Twitter. Our transformer-based classifier achieves an 85.06% macro F1 baseline score in detecting arguments. Moreover, our data reveals that Twitter users tend to engage in discussions involving informed inferences and information. TACO serves multiple purposes, such as training tweet classifiers to manage tweets based on inference and information elements, while also providing valuable insights into the conversational reply patterns of tweets.

The widespread online communication in a modern multilingual world has provided opportunities to blend more than one language (aka code-mixed language) in a single utterance. This has resulted a formidable challenge for the computational models due to the scarcity of annotated data and presence of noise. A potential solution to mitigate the data scarcity problem in low-resource setup is to leverage existing data in resource-rich language through translation. In this paper, we tackle the problem of code-mixed (Hinglish and Bengalish) to English machine translation. First, we synthetically develop HINMIX, a parallel corpus of Hinglish to English, with ~4.2M sentence pairs. Subsequently, we propose RCMT, a robust perturbation based joint-training model that learns to handle noise in the real-world code-mixed text by parameter sharing across clean and noisy words. Further, we show the adaptability of RCMT in a zero-shot setup for Bengalish to English translation. Our evaluation and comprehensive analyses qualitatively and quantitatively demonstrate the superiority of RCMT over state-of-the-art code-mixed and robust translation methods.

Joint entity-relation extraction remains a challenging task in information retrieval, given the intrinsic difficulty in modelling the interdependence between named entity recognition (NER) and relation extraction (RE) sub-tasks. Most existing joint extraction models encode entity and relation features in a sequential or parallel manner, allowing for limited one-way interaction. However, it is not yet clear how to capture the interdependence between these two sub-tasks in a synergistic and mutually reinforcing fashion. With this in mind, we propose a novel approach for joint entity-relation extraction, named Synergetic Interaction Network (SINET) which utilizes a cross-task attention mechanism to effectively leverage contextual associations between NER and RE. Specifically, we construct two sets of distinct token representations for NER and RE sub-tasks respectively. Then, both sets of unique representation interact with one another via a cross-task attention mechanism, which exploits associated contextual information produced by concerted efforts of both NER and RE. Experiments on three benchmark datasets demonstrate that the proposed model achieves significantly better performance in joint entity-relation extraction. Moreover, extended analysis validates that the proposed mechanism can indeed leverage the semantic information produced by NER and RE sub-tasks to boost one another in a complementary way. The source code is available to the public online.

Task-oriented dialogue (TOD) systems are introduced to solve specific tasks, which focus on training multiple tasks such as language understanding, tracking states, and generating appropriate responses to help users achieve their specific goals. Currently, one of the remaining challenges in this emergent research field is the capability to produce more robust architectures fine-tuned for end-to-end TOD systems. In this study, we consider this issue by exploiting the ability of pre-trained models to provide synthesis responses, which are then used as the input for the fine-tuned process. The main idea is to overcome the gap between the training process and inference process during fine-tuning end-to-end TOD systems. The experiment on Multiwoz datasets shows the effectiveness of our model compared with strong baselines in this research field. The source code is available for further exploitation.

Although there have been some works using prompt learning for the Aspect-based Sentiment Analysis(ABSA) tasks, their methods of prompt-tuning are simple and crude. Compared with vanilla fine-tuning methods, prompt learning intuitively bridges the objective form gap between pre-training and fine-tuning. Concretely, simply constructing prompt related to aspect words fails to fully exploit the potential of Pre-trained Language Models, and conducting more robust and professional prompt engineering for downstream tasks is a challenging problem that needs to be solved urgently. Therefore, in this paper, we propose a novel Syntax-aware Enhanced Prompt method (SynPrompt), which sufficiently mines the key syntactic information related to aspect words from the syntactic dependency tree. Additionally, to effectively harness the domain-specific knowledge embedded within PLMs for the ABSA tasks, we construct two adaptive prompt frameworks to enhance the perception ability of the above method. After conducting extensive experiments on three benchmark datasets, we have found that our method consistently achieves favorable results. These findings not only demonstrate the effectiveness and rationality of our proposed methods but also provide a powerful alternative to traditional prompt-tuning.

In this work, we introduce SwissSLi, the first sign language corpus that contains parallel data of all three Swiss sign languages, namely Swiss German Sign Language (DSGS), French Sign Language of Switzerland (LSF-CH), and Italian Sign Language of Switzerland (LIS-CH). The data underlying this corpus originates from television programs in three spoken languages: German, French, and Italian. The programs have for the most part been translated into sign language by deaf translators, resulting in a unique, up to six-way multi-parallel dataset between spoken and sign languages. We describe and release the sign language videos and spoken language subtitles as well as the overall statistics and some derivatives of the raw material. These derived components include cropped videos, pose estimation, phrase/sign-segmented videos, and sentence-segmented subtitles, all of which facilitate downstream tasks such as sign language transcription (glossing) and machine translation. The corpus is publicly available on the SWISSUbase data platform for research purposes only under a CC BY-NC-SA 4.0 license.