Enhancing Medical Multimodal Capabilities through PubMedVision: A Large-Scale, High-Quality Dataset for Injecting Visual Knowledge into Language Models
Constructing PubMedVision, a large-scale, high-quality medical multimodal dataset, to significantly boost the multimodal capabilities of language models in medical applications.
Evaluating the Potential of GPT-3.5 in Generating and Coding Discharge Summaries for Data Augmentation
GPT-3.5 can generate synthetic discharge summaries that, when combined with real data, improve the performance of local neural models on rare ICD-10 codes, but the generated documents lack the authenticity and narrative coherence required for clinical use.
HuatuoGPT-II: A Specialized Chinese Medical Language Model Developed through One-stage Training
A simplified one-stage domain adaptation protocol can effectively train a specialized Chinese medical language model, HuatuoGPT-II, that outperforms general language models and rivals proprietary models in the medical domain.
Evaluating Open-Source Large Language Models for Interpreting Pediatric Hypertension Guidelines
This research evaluates the performance of four open-source large language models (Meditron, MedAlpaca, Mistral, and Llama-2) in interpreting the European Society of Cardiology's pediatric hypertension guidelines.
Improving Medical Text Readability: A Comprehensive Study on Sentence-Level Complexity and Jargon
Properly measuring the readability of medical texts is crucial for making them more accessible to the general public. This study presents a systematic analysis of factors that contribute to the complexity of medical sentences, with a focus on the impact of specialized terminology and jargon.
Hippocrates: An Open-Source Framework for Advancing Large Language Models in Healthcare
Hippocrates is an open-source framework that aims to elevate the proficiency of large language models in medical reasoning and decision-making through continued pre-training, supervised fine-tuning, and reinforcement learning from AI-generated feedback.
Investigating Error Types in GPT-4 Responses to United States Medical Licensing Examination (USMLE) Questions
GPT-4 demonstrates high accuracy in answering USMLE questions, but around 14% of its responses contain errors. This study introduces a fine-grained error taxonomy to analyze these errors and provides a multi-label dataset of 300 annotated GPT-4 responses, along with associated medical concepts and semantic predications.
Qilin-Med: A Multi-stage Knowledge Injection Approach for Developing an Advanced Chinese Medical Language Model
A multi-stage training pipeline combining domain-specific Continued Pre-training, Supervised Fine-tuning, and Direct Preference Optimization can effectively transform a general-purpose language model into a specialized medical expert proficient in understanding complex medical texts and handling intricate medical tasks.
Medical mT5: An Open-Source Multilingual Text-to-Text Language Model for the Medical Domain
Medical mT5 is the first open-source multilingual text-to-text language model adapted to the medical domain, outperforming similarly-sized models in multi-task and zero-shot cross-lingual settings.
Automated Extraction and Analysis of Argumentative Structures in Medical Literature
Argument mining can help structure unstructured medical textual data by identifying argumentative components and their relations. This work investigates effective crosslingual transfer techniques to perform argument mining in medical texts for a target language (Spanish) when no annotated data is available, demonstrating the superiority of data-transfer over model-transfer approaches.
