Chain-of-Discussion: A Multi-Model Framework for Complex Evidence-Based Question Answering

Open-ended question answering requires models to find appropriate evidence to form well-reasoned, comprehensive and helpful answers. In practical applications, models also need to engage in extended discussions on potential scenarios closely relevant to the question. With augmentation of retrieval module, open-source Large Language Models (LLMs) can produce coherent answers often with different focuses, but are still sub-optimal in terms of reliable evidence selection and in-depth question analysis. In this paper, we propose a novel Chain-of-Discussion framework to leverage the synergy among multiple open-source LLMs aiming to provide \textbf{more correct} and \textbf{more comprehensive} answers for open-ended QA, although they are not strong enough individually. Our experiments show that discussions among multiple LLMs play a vital role in enhancing the quality of answers. We release our data and code at \url{https://github.com/kobayashikanna01/Chain-of-Discussion}.

CASA: Causality-driven Argument Sufficiency Assessment

The argument sufficiency assessment task aims to determine if the premises of a given argument support its conclusion. To tackle this task, existing works often train a classifier on data annotated by humans. However, annotating data is laborious, and annotations are often inconsistent due to subjective criteria. Motivated by the probability of sufficiency (PS) definition in the causal literature, we propose CASA, a zero-shot causality-driven argument sufficiency assessment framework. PS measures how likely introducing the premise event would lead to the conclusion, when both the premise and conclusion events are absent. To estimate this probability, we propose to use large language models (LLMs) to generate contexts that are inconsistent with the premise and conclusion, and revise them by injecting the premise event. Experiments on two logical fallacy detection datasets demonstrate that CASA accurately identifies insufficient arguments. We further deploy CASA in a writing assistance application, and find that suggestions generated by CASA enhance the sufficiency of student-written arguments. Code and data are available at https://github.com/xxxiaol/CASA.

MC^2: A Multilingual Corpus of Minority Languages in China

Large-scale corpora play a vital role in the construction of large language models (LLMs). However, existing LLMs exhibit limited abilities in understanding low-resource languages, including the minority languages in China, due to a lack of training data. To improve the accessibility of these languages, we present MC^2, a Multilingual Corpus of Minority Languages in China, which is the largest open-source corpus so far. It encompasses four underrepresented languages, i.e., Tibetan, Uyghur, Kazakh in the Kazakh Arabic script, and Mongolian in the traditional Mongolian script. Notably, two writing systems in MC^2 are long neglected in previous corpora. As we identify serious contamination in the low-resource language split in the existing multilingual corpora, we propose a quality-centric solution for collecting MC^2, prioritizing quality and accuracy while enhancing representativeness and diversity. By in-depth analysis, we demonstrate the new research challenges MC^2 brings, such as long-text modeling and multiplicity of writing systems. We hope MC^2 can help enhance the equity of the underrepresented languages in China and provide a reliable data foundation for further research on low-resource languages.

From the One, Judge of the Whole: Typed Entailment Graph Construction with Predicate Generation

Entailment Graphs (EGs) have been constructed based on extracted corpora as a strong and explainable form to indicate context-independent entailment relations in natural languages. However, EGs built by previous methods often suffer from the severe sparsity issues, due to limited corpora available and the long-tail phenomenon of predicate distributions. In this paper, we propose a multi-stage method, Typed Predicate-Entailment Graph Generator (TP-EGG), to tackle this problem. Given several seed predicates, TP-EGG builds the graphs by generating new predicates and detecting entailment relations among them. The generative nature of TP-EGG helps us leverage the recent advances from large pretrained language models (PLMs), while avoiding the reliance on carefully prepared corpora. Experiments on benchmark datasets show that TP-EGG can generate high-quality and scale-controllable entailment graphs, achieving significant in-domain improvement over state-of-the-art EGs and boosting the performance of down-stream inference tasks.

How Many Answers Should I Give? An Empirical Study of Multi-Answer Reading Comprehension

The multi-answer phenomenon, where a question may have multiple answers scattered in the document, can be well handled by humans but is challenging enough for machine reading comprehension (MRC) systems. Despite recent progress in multi-answer MRC, there lacks a systematic analysis of how this phenomenon arises and how to better address it. In this work, we design a taxonomy to categorize commonly-seen multi-answer MRC instances, with which we inspect three multi-answer datasets and analyze where the multi-answer challenge comes from. We further analyze how well different paradigms of current multi-answer MRC models deal with different types of multi-answer instances. We find that some paradigms capture well the key information in the questions while others better model the relationship between questions and contexts. We thus explore strategies to make the best of the strengths of different paradigms. Experiments show that generation models can be a promising platform to incorporate different paradigms. Our annotations and code are released for further research.

The Magic of IF: Investigating Causal Reasoning Abilities in Large Language Models of Code

Causal reasoning, the ability to identify cause-and-effect relationship, is crucial in human thinking. Although large language models (LLMs) succeed in many NLP tasks, it is still challenging for them to conduct complex causal reasoning like abductive reasoning and counterfactual reasoning. Given the fact that programming code may express causal relations more often and explicitly with conditional statements like ``if``, we want to explore whether Code-LLMs acquire better causal reasoning abilities. Our experiments show that compared to text-only LLMs, Code-LLMs with code prompts are significantly better in causal reasoning. We further intervene on the prompts from different aspects, and discover that the programming structure is crucial in code prompt design, while Code-LLMs are robust towards format perturbations.

More than Classification: A Unified Framework for Event Temporal Relation Extraction

Event temporal relation extraction~(ETRE) is usually formulated as a multi-label classification task, where each type of relation is simply treated as a one-hot label. This formulation ignores the meaning of relations and wipes out their intrinsic dependency. After examining the relation definitions in various ETRE tasks, we observe that all relations can be interpreted using the start and end time points of events. For example, relation \textit{Includes} could be interpreted as event 1 starting no later than event 2 and ending no earlier than event 2. In this paper, we propose a unified event temporal relation extraction framework, which transforms temporal relations into logical expressions of time points and completes the ETRE by predicting the relations between certain time point pairs. Experiments on TB-Dense and MATRES show significant improvements over a strong baseline and outperform the state-of-the-art model by 0.3\% on both datasets. By representing all relations in a unified framework, we can leverage the relations with sufficient data to assist the learning of other relations, thus achieving stable improvement in low-data scenarios. When the relation definitions are changed, our method can quickly adapt to the new ones by simply modifying the logic expressions that map time points to new event relations. The code is released at \url{https://github.com/AndrewZhe/A-Unified-Framework-for-ETRE}.

Lawyer LLaMA Technical Report

Large Language Models (LLMs), like LLaMA, have exhibited remarkable performances across various tasks. Nevertheless, when deployed to specific domains such as law or medicine, the models still confront the challenge of a deficiency in domain-specific knowledge and an inadequate capability to leverage that knowledge to resolve domain-related problems. In this paper, we focus on the legal domain and explore how to inject domain knowledge during the continual training stage and how to design proper supervised finetune tasks to help the model tackle practical issues. Moreover, to alleviate the hallucination problem during model's generation, we add a retrieval module and extract relevant articles before the model answers any queries. Augmenting with the extracted evidence, our model could generate more reliable responses. We release our data and model at https://github.com/AndrewZhe/lawyer-llama.

A Frustratingly Easy Improvement for Position Embeddings via Random Padding

Position embeddings, encoding the positional relationships among tokens in text sequences, make great contributions to modeling local context features in Transformer-based pre-trained language models. However, in Extractive Question Answering, position embeddings trained with instances of varied context lengths may not perform well as we expect. Since the embeddings of rear positions are updated fewer times than the front position embeddings, the rear ones may not be properly trained. In this paper, we propose a simple but effective strategy, Random Padding, without any modifications to architectures of existing pre-trained language models. We adjust the token order of input sequences when fine-tuning, to balance the number of updating times of every position embedding. Experiments show that Random Padding can significantly improve model performance on the instances whose answers are located at rear positions, especially when models are trained on short contexts but evaluated on long contexts. Our code and data will be released for future research.

Can BERT Refrain from Forgetting on Sequential Tasks? A Probing Study

Large pre-trained language models help to achieve state of the art on a variety of natural language processing (NLP) tasks, nevertheless, they still suffer from forgetting when incrementally learning a sequence of tasks. To alleviate this problem, recent works enhance existing models by sparse experience replay and local adaption, which yield satisfactory performance. However, in this paper we find that pre-trained language models like BERT have a potential ability to learn sequentially, even without any sparse memory replay. To verify the ability of BERT to maintain old knowledge, we adopt and re-finetune single-layer probe networks with the parameters of BERT fixed. We investigate the models on two types of NLP tasks, text classification and extractive question answering. Our experiments reveal that BERT can actually generate high quality representations for previously learned tasks in a long term, under extremely sparse replay or even no replay. We further introduce a series of novel methods to interpret the mechanism of forgetting and how memory rehearsal plays a significant role in task incremental learning, which bridges the gap between our new discovery and previous studies about catastrophic forgetting.