This paper proposes a novel approach to semantic QA over EHRs by first identifying the most relevant FHIR resources for a user query (Task1) and subsequently answering the query based on these resources (Task2). We explore the performance of privately hosted, fine-tuned LLMs, evaluating them against benchmark models such as GPT-4 and GPT-4o.

https://arxiv.org/pdf/2501.13687
#NLP #Healthcare

In this work, we propose a RAG QA system for the medical domain. Our method involves generating requisite prior knowledge to facilitate the answering process and compressing retrieved passages autoregressively with this knowledge. This approach ensures alignment between the question intent and the compressed context.

https://arxiv.org/pdf/2501.13567
#NLP #Healthcare

In this study, we propose a novel generative question answering framework for the biomedical domain which explicitly learns and incorporates evidence analysis with small language models (SLMs). The framework describes an evidence map for each question and utilizes an SLM to derive the representation of the supportive evaluation.

https://arxiv.org/pdf/2501.12746
#NLP #Healthcare

In this work, we present a deployable, small-scale medical language model designed for long-chain reasoning in clinical tasks using a self-evolution paradigm. Starting with a seed dataset of around 8,000 instances spanning five domains and 16 datasets, we prompt a base policy model to perform Monte Carlo Tree Search (MCTS) to construct verifiable reasoning chains.

https://arxiv.org/pdf/2501.12051
#NLP #Healthcare

In this work, we introduce a novel LLM physician framework that adheres to the Evidence-Based Medicine (EBM) process, efficiently integrating retrieval mechanisms as well as the selection and reasoning processes of evidence, thereby enhancing the problem-solving capabilities of LLMs in healthcare scenarios.

https://arxiv.org/pdf/2501.11885
#NLP #Healthcare

In this paper, we introduce a novel method termed Iterative Tree Analysis (ITA) for medical critics. ITA is designed to extract implicit claims from long medical texts and verify each claim through an iterative and adaptive tree-like reasoning process. This process involves a combination of top-down task decomposition and bottom-up evidence consolidation.

https://arxiv.org/pdf/2501.10642
#NLP #Healthcare

In this work, we propose a simple yet effective methodology to improve classification performance of images by LVLM, where we find concepts presents in the image by prompting the LVLM one concept at a time. Then, we ask the LVLM to classify the image based on the previous concept predictions.

https://arxiv.org/pdf/2501.12266
#NLP #Healthcare

This paper investigates the challenges of developing LLMs proficient in both multilingual understanding and medical knowledge. We demonstrate that simply translating medical data does not guarantee strong performance on clinical tasks in the target language, that the optimal language mix varies across tasks, and fine-tuning is not effective for incorporating new languages.

https://arxiv.org/pdf/2501.09825
#NLP #Healthcare

In this paper, we extract real patient interaction strategies from authentic doctor-patient conversations and use these strategies to guide the training of a patient simulator that closely mirrors real-world behavior. By inputting medical records into our patient simulator, we conduct extensive experiments to explore the relationship between "inquiry" and "diagnosis".

https://arxiv.org/pdf/2501.09484
#NLP #Healthcare

his study presents the first investigation into the abilities of LLMs to comprehend EHRs for patient data extraction and retrieval. We conduct extensive experiments using the MIMICSQL dataset to explore the impact of the prompt structure, instruction, context, and demonstration, of two backbone LLMs.

https://arxiv.org/pdf/2501.09384
#NLP #Healthcare