A5055530974- Topic Modeling
- Machine Learning in Healthcare
- Biomedical Text Mining and Ontologies
- Data Quality and Management
- Scientific Computing and Data Management
- Diabetic Foot Ulcer Assessment and Management
- Natural Language Processing Techniques
- COVID-19 diagnosis using AI
- Artificial Intelligence in Healthcare and Education
- Domain Adaptation and Few-Shot Learning
- Advanced Graph Neural Networks
- Bone fractures and treatments
- Multimodal Machine Learning Applications
- Lower Extremity Biomechanics and Pathologies
Korea Advanced Institute of Science and Technology
2023-2025
Kootenay Association for Science & Technology
2023
The development of large language models tailored for handling patients' clinical notes is often hindered by the limited accessibility and usability these due to strict privacy regulations. To address challenges, we first create synthetic large-scale using publicly available case reports extracted from biomedical literature. We then use train our specialized model, Asclepius. While Asclepius trained on data, assess its potential performance in real-world applications evaluating it real...
Electronic Health Records (EHRs), which contain patients' medical histories in various multi-modal formats, often overlook the potential for joint reasoning across imaging and table modalities underexplored current EHR Question Answering (QA) systems. In this paper, we introduce EHRXQA, a novel question answering dataset combining structured EHRs chest X-ray images. To develop our dataset, first construct two uni-modal resources: 1) The MIMIC-CXR-VQA newly created visual (VQA) benchmark,...
Gangwoo Kim, Hajung Lei Ji, Seongsu Bae, Chanhwi Mujeen Sung, Hyunjae Kun Yan, Eric Chang, Jaewoo Kang. The 22nd Workshop on Biomedical Natural Language Processing and BioNLP Shared Tasks. 2023.
Electronic Health Records (EHRs) are relational databases that store the entire medical histories of patients within hospitals. They record numerous aspects patients' care, from hospital admission and diagnosis to treatment discharge. While EHRs vital sources clinical data, exploring them beyond a predefined set queries requires skills in query languages like SQL. To make information retrieval more accessible, one strategy is build question-answering system, possibly leveraging text-to-SQL...
Electronic Health Records (EHRs) are integral for storing comprehensive patient medical records, combining structured data (e.g., medications) with detailed clinical notes physician notes). These elements essential straightforward retrieval and provide deep, contextual insights into care. However, they often suffer from discrepancies due to unintuitive EHR system designs human errors, posing serious risks safety. To address this, we developed EHRCon, a new dataset task specifically designed...
Question Answering on Electronic Health Records (EHR-QA) has a significant impact the healthcare domain, and it is being actively studied. Previous research structured EHR-QA focuses converting natural language queries into query such as SQL or SPARQL (NLQ2Query), so problem scope limited to pre-defined data types by specific language. In order expand task beyond this limitation handle multi-modal medical solve complex inference in future, more primitive systemic needed. paper, we design...
In this paper, we introduce CheXOFA, a new pre-trained vision-language model (VLM) for the chest X-ray domain. Our is initially on various multimodal datasets within general domain before being transferred to Following prominent VLM, unify domain-specific tasks into simple sequence-to-sequence schema. It enables effectively learn required knowledge and skills from limited resources in Demonstrating superior performance benchmark provided by BioNLP shared task, our benefits its training...