Better Explain Transformers by Illuminating Important Information

Transformer-based models excel in various natural language processing (NLP) tasks, attracting countless efforts to explain their inner workings. Prior methods explain Transformers by focusing on the raw gradient and attention as token attribution scores, where non-relevant information is often considered during explanation computation, resulting in confusing results. In this work, we propose highlighting the important information and eliminating irrelevant information by a refined information flow on top of the layer-wise relevance propagation (LRP) method. Specifically, we consider identifying syntactic and positional heads as important attention heads and focus on the relevance obtained from these important heads. Experimental results demonstrate that irrelevant information does distort output attribution scores and then should be masked during explanation computation. Compared to eight baselines on both classification and question-answering datasets, our method consistently outperforms with over 3\% to 33\% improvement on explanation metrics, providing superior explanation performance. Our anonymous code repository is available at: https://github.com/LinxinS97/Mask-LRP

RecPrompt: A Prompt Tuning Framework for News Recommendation Using Large Language Models

In the evolving field of personalized news recommendation, understanding the semantics of the underlying data is crucial. Large Language Models (LLMs) like GPT-4 have shown promising performance in understanding natural language. However, the extent of their applicability in news recommendation systems remains to be validated. This paper introduces RecPrompt, the first framework for news recommendation that leverages the capabilities of LLMs through prompt engineering. This system incorporates a prompt optimizer that applies an iterative bootstrapping process, enhancing the LLM-based recommender's ability to align news content with user preferences and interests more effectively. Moreover, this study offers insights into the effective use of LLMs in news recommendation, emphasizing both the advantages and the challenges of incorporating LLMs into recommendation systems.

MedGen: A Python Natural Language Processing Toolkit for Medical Text Processing

This study introduces MedGen, a comprehensive natural language processing (NLP) toolkit designed for medical text processing. MedGen is tailored for biomedical researchers and healthcare professionals with an easy-to-use, all-in-one solution that requires minimal programming expertise. It includes (1) Generative Functions: For the first time, MedGen includes four advanced generative functions: question answering, text summarization, text simplification, and machine translation; (2) Basic NLP Functions: MedGen integrates 12 essential NLP functions such as word tokenization and sentence segmentation; and (3) Query and Search Capabilities: MedGen provides user-friendly query and search functions on text corpora. We fine-tuned 32 domain-specific language models, evaluated them thoroughly on 24 established benchmarks and conducted manual reviews with clinicians. Additionally, we expanded our toolkit by introducing query and search functions, while also standardizing and integrating functions from third-party libraries. The toolkit, its models, and associated data are publicly available via https://github.com/Yale-LILY/MedGen.

Integrating UMLS Knowledge into Large Language Models for Medical Question Answering

Large language models (LLMs) have demonstrated powerful text generation capabilities, bringing unprecedented innovation to the healthcare field. While LLMs hold immense promise for applications in healthcare, applying them to real clinical scenarios presents significant challenges, as these models may generate content that deviates from established medical facts and even exhibit potential biases. In our research, we develop an augmented LLM framework based on the Unified Medical Language System (UMLS), aiming to better serve the healthcare community. We employ LLaMa2-13b-chat and ChatGPT-3.5 as our benchmark models, and conduct automatic evaluations using the ROUGE Score and BERTScore on 104 questions from the LiveQA test set. Additionally, we establish criteria for physician-evaluation based on four dimensions: Factuality, Completeness, Readability and Relevancy. ChatGPT-3.5 is used for physician evaluation with 20 questions on the LiveQA test set. Multiple resident physicians conducted blind reviews to evaluate the generated content, and the results indicate that this framework effectively enhances the factuality, completeness, and relevance of generated content. Our research demonstrates the effectiveness of using UMLS-augmented LLMs and highlights the potential application value of LLMs in in medical question-answering.

NLPBench: Evaluating Large Language Models on Solving NLP Problems

Recent developments in large language models (LLMs) have shown promise in enhancing the capabilities of natural language processing (NLP). Despite these successes, there remains a dearth of research dedicated to the NLP problem-solving abilities of LLMs. To fill the gap in this area, we present a unique benchmarking dataset, NLPBench, comprising 378 college-level NLP questions spanning various NLP topics sourced from Yale University's prior final exams. NLPBench includes questions with context, in which multiple sub-questions share the same public information, and diverse question types, including multiple choice, short answer, and math. Our evaluation, centered on LLMs such as GPT-3.5/4, PaLM-2, and LLAMA-2, incorporates advanced prompting strategies like the chain-of-thought (CoT) and tree-of-thought (ToT). Our study reveals that the effectiveness of the advanced prompting strategies can be inconsistent, occasionally damaging LLM performance, especially in smaller models like the LLAMA-2 (13b). Furthermore, our manual assessment illuminated specific shortcomings in LLMs' scientific problem-solving skills, with weaknesses in logical decomposition and reasoning notably affecting results.

Large Language Models on Wikipedia-Style Survey Generation: an Evaluation in NLP Concepts

Large Language Models (LLMs) have achieved significant success across various natural language processing (NLP) tasks, encompassing question-answering, summarization, and machine translation, among others. While LLMs excel in general tasks, their efficacy in domain-specific applications remains under exploration. Additionally, LLM-generated text sometimes exhibits issues like hallucination and disinformation. In this study, we assess LLMs' capability of producing concise survey articles within the computer science-NLP domain, focusing on 20 chosen topics. Automated evaluations indicate that GPT-4 outperforms GPT-3.5 when benchmarked against the ground truth. Furthermore, four human evaluators provide insights from six perspectives across four model configurations. Through case studies, we demonstrate that while GPT often yields commendable results, there are instances of shortcomings, such as incomplete information and the exhibition of lapses in factual accuracy.

Going Beyond Local: Global Graph-Enhanced Personalized News Recommendations

Precisely recommending candidate news articles to users has always been a core challenge for personalized news recommendation systems. Most recent works primarily focus on using advanced natural language processing techniques to extract semantic information from rich textual data, employing content-based methods derived from local historical news. However, this approach lacks a global perspective, failing to account for users' hidden motivations and behaviors beyond semantic information. To address this challenge, we propose a novel model called GLORY (Global-LOcal news Recommendation sYstem), which combines global representations learned from other users with local representations to enhance personalized recommendation systems. We accomplish this by constructing a Global-aware Historical News Encoder, which includes a global news graph and employs gated graph neural networks to enrich news representations, thereby fusing historical news representations by a historical news aggregator. Similarly, we extend this approach to a Global Candidate News Encoder, utilizing a global entity graph and a candidate news aggregator to enhance candidate news representation. Evaluation results on two public news datasets demonstrate that our method outperforms existing approaches. Furthermore, our model offers more diverse recommendations.

Topic-Centric Explanations for News Recommendation

News recommender systems (NRS) have been widely applied for online news websites to help users find relevant articles based on their interests. Recent methods have demonstrated considerable success in terms of recommendation performance. However, the lack of explanation for these recommendations can lead to mistrust among users and lack of acceptance of recommendations. To address this issue, we propose a new explainable news model to construct a topic-aware explainable recommendation approach that can both accurately identify relevant articles and explain why they have been recommended, using information from associated topics. Additionally, our model incorporates two coherence metrics applied to assess topic quality, providing measure of the interpretability of these explanations. The results of our experiments on the MIND dataset indicate that the proposed explainable NRS outperforms several other baseline systems, while it is also capable of producing interpretable topics compared to those generated by a classical LDA topic model. Furthermore, we present a case study through a real-world example showcasing the usefulness of our NRS for generating explanations.

HiPool: Modeling Long Documents Using Graph Neural Networks

Encoding long sequences in Natural Language Processing (NLP) is a challenging problem. Though recent pretraining language models achieve satisfying performances in many NLP tasks, they are still restricted by a pre-defined maximum length, making them challenging to be extended to longer sequences. So some recent works utilize hierarchies to model long sequences. However, most of them apply sequential models for upper hierarchies, suffering from long dependency issues. In this paper, we alleviate these issues through a graph-based method. We first chunk the sequence with a fixed length to model the sentence-level information. We then leverage graphs to model intra- and cross-sentence correlations with a new attention mechanism. Additionally, due to limited standard benchmarks for long document classification (LDC), we propose a new challenging benchmark, totaling six datasets with up to 53k samples and 4034 average tokens' length. Evaluation shows our model surpasses competitive baselines by 2.6% in F1 score, and 4.8% on the longest sequence dataset. Our method is shown to outperform hierarchical sequential models with better performance and scalability, especially for longer sequences.

EHRKit: A Python Natural Language Processing Toolkit for Electronic Health Record Texts

The Electronic Health Record (EHR) is an essential part of the modern medical system and impacts healthcare delivery, operations, and research. Unstructured text is attracting much attention despite structured information in the EHRs and has become an exciting research field. The success of the recent neural Natural Language Processing (NLP) method has led to a new direction for processing unstructured clinical notes. In this work, we create a python library for clinical texts, EHRKit. This library contains two main parts: MIMIC-III-specific functions and tasks specific functions. The first part introduces a list of interfaces for accessing MIMIC-III NOTEEVENTS data, including basic search, information retrieval, and information extraction. The second part integrates many third-party libraries for up to 12 off-shelf NLP tasks such as named entity recognition, summarization, machine translation, etc.