MARTI: A Framework for Multi-Agent LLM Systems Reinforced Training and Inference
Introduces an open framework for reinforced training and inference in multi-agent LLM systems.
Publications
This selection emphasizes research that moves between methodological innovation and real healthcare relevance. For the latest updates, see Google Scholar.
Archive
Grouped by year for a clean reading flow rather than a legacy CV dump.
Frames longitudinal risk prediction as a coordinated agent workflow over patient trajectory evidence.
Learns smooth latent patient trajectories that improve survival prediction while preserving interpretability.
Explores reinforcement learning for more grounded, analytically useful automated scientific reviews.
Connects Whole Health services to measurable tobacco cessation outcomes in veterans.
Examines population-level tobacco cessation outcomes associated with Whole Health implementation in the Veterans Health Administration.
Builds broadly useful biomedical models without sacrificing domain specialization.
Evaluates how well large language models can support biomedical hypothesis generation.
Uses explanation-augmented contrastive learning to improve biomedical term representations and large-scale knowledge graph construction.
Investigates whether large language models can propose scientifically meaningful biomedical hypotheses without task-specific training.
Introduces hierarchy-aware pretraining for biomedical term embeddings to better encode graded semantic relatedness.
Presents a large-scale biomedical knowledge graph generated algorithmically rather than through traditional manual curation.
Introduces fine-grained biomedical term representations for improved clustering and knowledge organization.
Studies learning-based 2-D microwave thorax imaging with the supervised descent method for fast structural reconstruction.