Improving Factuality in Biomedical Long-form Question Answering
Introducing OLAPH, a novel framework that leverages cost-effective and multifaceted automatic evaluation to generate synthetic preference sets and optimize long-form answers from large language models to be more factual and coherent in the biomedical domain.
Comprehensive Evaluation of Large Language Models for Biomedical Natural Language Processing: Benchmarks, Baselines, and Recommendations
State-of-the-art fine-tuning approaches outperformed zero- and few-shot large language models in most biomedical NLP tasks, but closed-source LLMs like GPT-3.5 and GPT-4 achieved better performance in reasoning-related tasks and competitive accuracy in generation-related tasks.
Enhancing Biomedical Language Models with Retrieval-Augmented Generation: The BiomedRAG Framework
The BiomedRAG framework effectively integrates retrieved chunk-based documents into large language models to enhance their performance across various biomedical NLP tasks, including information extraction, text classification, link prediction, and question answering.
Extracting Fine-Grained Experimental Findings from Biomedical Literature
This work presents CARE, a new dataset and annotation schema for extracting fine-grained experimental findings from biomedical literature, including clinical trials and case reports.
Enhancing Large Language Models for Biomedical Named Entity Recognition through Prompt Engineering and Knowledge Integration
Leveraging prompt engineering, strategic in-context example selection, and external knowledge integration, this study demonstrates significant improvements in the performance of large language models for biomedical named entity recognition tasks.
Safe and Reliable Biomedical Natural Language Inference for Clinical Trials
Developing robust and dependable natural language inference (NLI) models for clinical trial data to support safer and more trustworthy AI assistance in healthcare decision-making.
Evaluating Large Language Models' Capabilities for Safe Biomedical Natural Language Inference on Clinical Trial Reports
Large language models (LLMs) can achieve strong performance on natural language inference tasks in the biomedical domain, but they still face challenges in maintaining consistency, faithfulness, and robust reasoning, especially when dealing with numerical and logical reasoning on clinical trial reports.
Evaluating Masked and Generative Language Models on Natural Language Inference for Clinical Trial Data
Comparing the performance of masked language models and generative language models on a natural language inference task for clinical trial data, focusing on metrics of faithfulness and consistency.
Comprehensive Evaluation of Intent Detection and Named Entity Recognition in Biomedical Literature
Supervised Fine-Tuned approaches, such as RoBERTa and BINDER (PubMedBERT), outperform general-purpose Large Language Models like ChatGPT on intent detection and named entity recognition tasks in the biomedical domain.
Enhancing Named Entity Recognition in the Chemical Domain via Metric Learning and Pseudo-Labeling
A model is proposed to effectively transfer knowledge from the biomedical domain to the chemical domain for named entity recognition, by projecting source and target entities into separate regions of the feature space.
