Structured medical question answering work combining RDF retrieval with claim-level verification for faithful generative AI outputs. Contribution: proposed claim-level verification as the core faithfulness mechanism, designed the structured retrieval pipeline, and led evaluation design.
@article{natalias2026structure,title={Structure-Grounded Medical QA: RDF Retrieval and Claim-Level Verification for Faithful Answering},author={Natalias, Stocks and Salve, Jatin Avinash and Hunter and Kunchepu and Herrera and Youm and Gilda and Dorr},journal={ACL 2026 Workshop on Structured Understanding and Reasoning for Generative LLMs (SURGeLLM)},year={2026},note={Submitted}}
UF
Anisotropic Noise Injection for Improving Utility in Differentially Private SGD
Differential privacy work on shaping injected noise along gradient covariance eigenvectors to improve the utility of private SGD. Contribution: originated the core idea and led theoretical derivation and empirical evaluation.
@article{salve2026anisotropic,title={Anisotropic Noise Injection for Improving Utility in Differentially Private SGD},author={Salve, Jatin Avinash and Nahar and Mali},journal={University of Florida},year={2026},note={Under review}}
2024
ACL
From Sights to Insights: Towards Summarization of Multimodal Clinical Documents
Soumya Ghosh, Mohit Tomar, Anshul Tiwari, and 3 more authors
In Proceedings of the 62nd Annual Meeting of the Association for Computational LinguisticsACL 2024 long paper, main conference. , Aug 2024
Multimodal clinical document summarization work focused on grounding generated summaries in visual and textual evidence. Contribution: designed the vision cross-attention fusion module and ran ablation studies isolating the contribution of multimodal grounding.
@inproceedings{ghosh2024sights,title={From Sights to Insights: Towards Summarization of Multimodal Clinical Documents},author={Ghosh, Soumya and Tomar, Mohit and Tiwari, Anshul and Saha, Sriparna and Salve, Jatin Avinash and Sinha, Manjira},booktitle={Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics},year={2024},month=aug,publisher={Association for Computational Linguistics,},}