Computation and Language (cs.CL)

Abstract: Link prediction task is vital to automatically understanding the structure of large knowledge bases. In this paper, we present our system to solve this task at the Data Science and Advanced Analytics 2023 Competition "Efficient and Effective Link Prediction" (DSAA-2023 Competition) with a corpus containing 948,233 training and 238,265 for public testing. This paper introduces an approach to link prediction in Wikipedia articles by formulating it as a natural language inference (NLI) task. Drawing inspiration from recent advancements in natural language processing and understanding, we cast link prediction as an NLI task, wherein the presence of a link between two articles is treated as a premise, and the task is to determine whether this premise holds based on the information presented in the articles. We implemented our system based on the Sentence Pair Classification for Link Prediction for the Wikipedia Articles task. Our system achieved 0.99996 Macro F1-score and 1.00000 Macro F1-score for the public and private test sets, respectively. Our team UIT-NLP ranked 3rd in performance on the private test set, equal to the scores of the first and second places. Our code is publicly for research purposes.

Abstract: Multi-modal Large Language Model (MLLM) refers to a model expanded from a Large Language Model (LLM) that possesses the capability to handle and infer multi-modal data. Current MLLMs typically begin by using LLMs to decompose tasks into multiple subtasks, then employing individual pre-trained models to complete specific subtasks, and ultimately utilizing LLMs to integrate the results of each subtasks to obtain the results of the task. In real-world scenarios, when dealing with large projects, it is common practice to break down the project into smaller sub-projects, with different teams providing corresponding solutions or results. The project owner then decides which solution or result to use, ensuring the best possible outcome for each subtask and, consequently, for the entire project. Inspired by this, this study considers selecting multiple pre-trained models to complete the same subtask. By combining the results from multiple pre-trained models, the optimal subtask result is obtained, enhancing the performance of the MLLM. Specifically, this study first selects multiple pre-trained models focused on the same subtask based on distinct evaluation approaches, and then invokes these models in parallel to process input data and generate corresponding subtask results. Finally, the results from multiple pre-trained models for the same subtask are compared using the LLM, and the best result is chosen as the outcome for that subtask. Extensive experiments are conducted in this study using GPT-4 annotated datasets and human-annotated datasets. The results of various evaluation metrics adequately demonstrate the effectiveness of the proposed approach in this paper.

Abstract: We propose TCSP, a novel method for compressing a transformer model by focusing on reducing the hidden size of the model. By projecting the whole transform model into a subspace, we enable matrix operations between the weight matrices in the model and features in a reduced-dimensional space, leading to significant reductions in model parameters and computing resources. To establish this subspace, we decompose the feature matrix, derived from different layers of sampled data instances, into a projection matrix. For evaluation, TCSP is applied to compress T5 and BERT models on the GLUE and SQuAD benchmarks. Experimental results demonstrate that TCSP achieves a compression ratio of 44\% with at most 1.6\% degradation in accuracy, surpassing or matching prior compression methods. Furthermore, TCSP exhibits compatibility with other methods targeting filter and attention head size compression.

Abstract: Ambiguities in natural language give rise to probability distributions over interpretations. The distributions are often over multiple ambiguous words at a time; a multiplicity which makes them a suitable topic for sheaf-theoretic models of quantum contextuality. Previous research showed that different quantitative measures of contextuality correlate well with Psycholinguistic research on lexical ambiguities. In this work, we focus on coreference ambiguities and investigate the Winograd Schema Challenge (WSC), a test proposed by Levesque in 2011 to evaluate the intelligence of machines. The WSC consists of a collection of multiple-choice questions that require disambiguating pronouns in sentences structured according to the Winograd schema, in a way that makes it difficult for machines to determine the correct referents but remains intuitive for human comprehension. In this study, we propose an approach that analogously models the Winograd schema as an experiment in quantum physics. However, we argue that the original Winograd Schema is inherently too simplistic to facilitate contextuality. We introduce a novel mechanism for generalising the schema, rendering it analogous to a Bell-CHSH measurement scenario. We report an instance of this generalised schema, complemented by the human judgements we gathered via a crowdsourcing platform. The resulting model violates the Bell-CHSH inequality by 0.192, thus exhibiting contextuality in a coreference resolution setting.

Abstract: This paper is a summary of the work in my PhD thesis. In which, I investigate the impact of bias in NLP models on the task of hate speech detection from three perspectives: explainability, offensive stereotyping bias, and fairness. I discuss the main takeaways from my thesis and how they can benefit the broader NLP community. Finally, I discuss important future research directions. The findings of my thesis suggest that bias in NLP models impacts the task of hate speech detection from all three perspectives. And that unless we start incorporating social sciences in studying bias in NLP models, we will not effectively overcome the current limitations of measuring and mitigating bias in NLP models.

Abstract: We present a general framework for unsupervised text style transfer with deep generative models. The framework models each sentence-label pair in the non-parallel corpus as partially observed from a complete quadruplet which additionally contains two latent codes representing the content and style, respectively. These codes are learned by exploiting dependencies inside the observed data. Then a sentence is transferred by manipulating them. Our framework is able to unify previous embedding and prototype methods as two special forms. It also provides a principled perspective to explain previously proposed techniques in the field such as aligned encoder and adversarial training. We further conduct experiments on three benchmarks. Both automatic and human evaluation results show that our methods achieve better or competitive results compared to several strong baselines.

Abstract: As the key to sentiment analysis, sentiment composition considers the classification of a constituent via classifications of its contained sub-constituents and rules operated on them. Such compositionality has been widely studied previously in the form of hierarchical trees including untagged and sentiment ones, which are intrinsically suboptimal in our view. To address this, we propose semantic tree, a new tree form capable of interpreting the sentiment composition in a principled way. Semantic tree is a derivation of a context-free grammar (CFG) describing the specific composition rules on difference semantic roles, which is designed carefully following previous linguistic conclusions. However, semantic tree is a latent variable since there is no its annotation in regular datasets. Thus, in our method, it is marginalized out via inside algorithm and learned to optimize the classification performance. Quantitative and qualitative results demonstrate that our method not only achieves better or competitive results compared to baselines in the setting of regular and domain adaptation classification, and also generates plausible tree explanations.

Abstract: We present DictaBERT, a new state-of-the-art pre-trained BERT model for modern Hebrew, outperforming existing models on most benchmarks. Additionally, we release two fine-tuned versions of the model, designed to perform two specific foundational tasks in the analysis of Hebrew texts: prefix segmentation and morphological tagging. These fine-tuned models allow any developer to perform prefix segmentation and morphological tagging of a Hebrew sentence with a single call to a HuggingFace model, without the need to integrate any additional libraries or code. In this paper we describe the details of the training as well and the results on the different benchmarks. We release the models to the community, along with sample code demonstrating their use. We release these models as part of our goal to help further research and development in Hebrew NLP.

Abstract: Purpose: In this paper, we present an automated method for article classification, leveraging the power of Large Language Models (LLM). The primary focus is on the field of ophthalmology, but the model is extendable to other fields. Methods: We have developed a model based on Natural Language Processing (NLP) techniques, including advanced LLMs, to process and analyze the textual content of scientific papers. Specifically, we have employed zero-shot learning (ZSL) LLM models and compared against Bidirectional and Auto-Regressive Transformers (BART) and its variants, and Bidirectional Encoder Representations from Transformers (BERT), and its variant such as distilBERT, SciBERT, PubmedBERT, BioBERT. Results: The classification results demonstrate the effectiveness of LLMs in categorizing large number of ophthalmology papers without human intervention. Results: To evalute the LLMs, we compiled a dataset (RenD) of 1000 ocular disease-related articles, which were expertly annotated by a panel of six specialists into 15 distinct categories. The model achieved mean accuracy of 0.86 and mean F1 of 0.85 based on the RenD dataset. Conclusion: The proposed framework achieves notable improvements in both accuracy and efficiency. Its application in the domain of ophthalmology showcases its potential for knowledge organization and retrieval in other domains too. We performed trend analysis that enables the researchers and clinicians to easily categorize and retrieve relevant papers, saving time and effort in literature review and information gathering as well as identification of emerging scientific trends within different disciplines. Moreover, the extendibility of the model to other scientific fields broadens its impact in facilitating research and trend analysis across diverse disciplines.

Abstract: Current speech large language models build upon discrete speech representations, which can be categorized into semantic tokens and acoustic tokens. However, existing speech tokens are not specifically designed for speech language modeling. To assess the suitability of speech tokens for building speech language models, we established the first benchmark, SLMTokBench. Our results indicate that neither semantic nor acoustic tokens are ideal for this purpose. Therefore, we propose SpeechTokenizer, a unified speech tokenizer for speech large language models. SpeechTokenizer adopts the Encoder-Decoder architecture with residual vector quantization (RVQ). Unifying semantic and acoustic tokens, SpeechTokenizer disentangles different aspects of speech information hierarchically across different RVQ layers. Furthermore, We construct a Unified Speech Language Model (USLM) leveraging SpeechTokenizer. Experiments show that SpeechTokenizer performs comparably to EnCodec in speech reconstruction and demonstrates strong performance on the SLMTokBench benchmark. Also, USLM outperforms VALL-E in zero-shot Text-to-Speech tasks. Code and models are available at https://github.com/ZhangXInFD/SpeechTokenizer/.

Abstract: English datasets predominantly reflect the perspectives of certain nationalities, which can lead to cultural biases in models and datasets. This is particularly problematic in tasks heavily influenced by subjectivity, such as hate speech detection. To delve into how individuals from different countries perceive hate speech, we introduce CReHate, a cross-cultural re-annotation of the sampled SBIC dataset. This dataset includes annotations from five distinct countries: Australia, Singapore, South Africa, the United Kingdom, and the United States. Our thorough statistical analysis highlights significant differences based on nationality, with only 59.4% of the samples achieving consensus among all countries. We also introduce a culturally sensitive hate speech classifier via transfer learning, adept at capturing perspectives of different nationalities. These findings underscore the need to re-evaluate certain aspects of NLP research, especially with regard to the nuanced nature of hate speech in the English language.

Abstract: Chain-of-Thought Prompting (CoT) reinforces the reasoning capabilities of Large Language Models (LLMs) through the generation of intermediate rationales. However, these enhancements predominantly benefit large-scale models, leaving small LMs without significant performance improvements when directly applying CoT. Despite the advanced reasoning capabilities of LLMs, CoT relies primarily on their pre-trained internal knowledge. The external knowledge that is previously unknown to the model remains unexploited. This omission becomes pronounced in tasks such as stance detection, where the external background knowledge plays a pivotal role. Additionally, the large-scale architecture of LLMs inevitably present efficiency challenges during deployment. To address these challenges, we introduce the Ladder-of-Thought (LoT) for stance detection. Grounded in a dual-phase Cascaded Optimization framework, LoT directs the model to incorporate high-quality external knowledge, enhancing the intermediate rationales it generates. These bolstered rationales subsequently serve as the foundation for more precise predictions - akin to how a ladder facilitates reaching elevated goals. LoT achieves a balance between efficiency and accuracy, making it an adaptable and efficient framework for stance detection. Our empirical evaluations underscore LoT's effectiveness, marking a 16% improvement over ChatGPT and a 10% enhancement compared to ChatGPT with CoT.

Abstract: The increased digitization of the labour market has given researchers, educators, and companies the means to analyze and better understand the labour market. However, labour market resources, although available in high volumes, tend to be unstructured, and as such, research towards methodologies for the identification, linking, and extraction of entities becomes more and more important. Against the backdrop of this quest for better labour market representations, resource constraints and the unavailability of large-scale annotated data cause a reliance on human domain experts. We demonstrate the effectiveness of prompt-based tuning of pre-trained language models (PLM) in labour market specific applications. Our results indicate that cost-efficient methods such as PTR and instruction tuning without exemplars can significantly increase the performance of PLMs on downstream labour market applications without introducing additional model layers, manual annotations, and data augmentation.

Abstract: The main limiting factor in the development of robust multilingual dialogue evaluation metrics is the lack of multilingual data and the limited availability of open sourced multilingual dialogue systems. In this work, we propose a workaround for this lack of data by leveraging a strong multilingual pretrained LLM and augmenting existing English dialogue data using Machine Translation. We empirically show that the naive approach of finetuning a pretrained multilingual encoder model with translated data is insufficient to outperform the strong baseline of finetuning a multilingual model with only source data. Instead, the best approach consists in the careful curation of translated data using MT Quality Estimation metrics, excluding low quality translations that hinder its performance.

Abstract: Despite significant research effort in the development of automatic dialogue evaluation metrics, little thought is given to evaluating dialogues other than in English. At the same time, ensuring metrics are invariant to semantically similar responses is also an overlooked topic. In order to achieve the desired properties of robustness and multilinguality for dialogue evaluation metrics, we propose a novel framework that takes advantage of the strengths of current evaluation models with the newly-established paradigm of prompting Large Language Models (LLMs). Empirical results show our framework achieves state of the art results in terms of mean Spearman correlation scores across several benchmarks and ranks first place on both the Robust and Multilingual tasks of the DSTC11 Track 4 "Automatic Evaluation Metrics for Open-Domain Dialogue Systems", proving the evaluation capabilities of prompted LLMs.

Abstract: When human programmers have mastered a programming language, it would be easier when they learn a new programming language. In this report, we focus on exploring whether programming languages can boost each other during the instruction fine-tuning phase of code large language models. We conduct extensive experiments of 8 popular programming languages (Python, JavaScript, TypeScript, C, C++, Java, Go, HTML) on StarCoder. Results demonstrate that programming languages can significantly improve each other. For example, CodeM-Python 15B trained on Python is able to increase Java by an absolute 17.95% pass@1 on HumanEval-X. More surprisingly, we found that CodeM-HTML 7B trained on the HTML corpus can improve Java by an absolute 15.24% pass@1. Our training data is released at https://github.com/NL2Code/CodeM.

Abstract: Gender biases in language generation systems are challenging to mitigate. One possible source for these biases is gender representation disparities in the training and evaluation data. Despite recent progress in documenting this problem and many attempts at mitigating it, we still lack shared methodology and tooling to report gender representation in large datasets. Such quantitative reporting will enable further mitigation, e.g., via data augmentation. This paper describes the Gender-GAP Pipeline (for Gender-Aware Polyglot Pipeline), an automatic pipeline to characterize gender representation in large-scale datasets for 55 languages. The pipeline uses a multilingual lexicon of gendered person-nouns to quantify the gender representation in text. We showcase it to report gender representation in WMT training data and development data for the News task, confirming that current data is skewed towards masculine representation. Having unbalanced datasets may indirectly optimize our systems towards outperforming one gender over the others. We suggest introducing our gender quantification pipeline in current datasets and, ideally, modifying them toward a balanced representation.

Abstract: We present Belebele, a multiple-choice machine reading comprehension (MRC) dataset spanning 122 language variants. Significantly expanding the language coverage of natural language understanding (NLU) benchmarks, this dataset enables the evaluation of text models in high-, medium-, and low-resource languages. Each question is based on a short passage from the Flores-200 dataset and has four multiple-choice answers. The questions were carefully curated to discriminate between models with different levels of general language comprehension. The English dataset on its own proves difficult enough to challenge state-of-the-art language models. Being fully parallel, this dataset enables direct comparison of model performance across all languages. We use this dataset to evaluate the capabilities of multilingual masked language models (MLMs) and large language models (LLMs). We present extensive results and find that despite significant cross-lingual transfer in English-centric LLMs, much smaller MLMs pretrained on balanced multilingual data still understand far more languages. We also observe that larger vocabulary size and conscious vocabulary construction correlate with better performance on low-resource languages. Overall, Belebele opens up new avenues for evaluating and analyzing the multilingual capabilities of NLP systems.