Avisha Das

Avisha Das

Research Fellow

Mayo Clinic Arizona

Biography

I am currently on the job market and looking for tenure-track faculty positions in Biomedical Informatics with a focus on Clinical NLP and Secure Large Language Modeling.

I am currently a Research Fellow with the Arizona Advanced AI & Innovation (A3I) Hub, Mayo Clinic, AZ. I have worked as a Postdoc at University of Texas Health Science Center, Houston with Dr. Wenjin J. Zheng and Dr. Hua Xu on Biomedical NLP and Knowledge Mining. I earned my PhD from University of Houston under Dr. Rakesh Verma.

My research interests lie in Large Language Modeling and Language Understanding. I am currently working on exploring LLM vulnerabilities in the Biomedical and Clinical Domain. I have previously worked on Biomedical Knowledge Mining and Retrieval as well as Empathetic Conversational Agents for Mental Health Support.

Research Areas: Generative Language Modeling, Information Retrieval, Biomedical NLP, Security Analytics

Interests
  • Natural Language Generation and Understanding
  • Security and Large Language Modeling
  • Knowledge Mining and Information Retrieval
Education
  • PhD, Computer Science, 2020

    University of Houston, Houston, TX

  • BTech, Electronics Engineering, 2014

    West Bengal University of Technology, Kolkata, IN

Experience

 
 
 
 
 
Research Fellow
Arizona Advanced AI & Innovation (A3I) Hub, Mayo Clinic Arizona
November 2023 – Present Phoenix, Arizona
  • Security and Clinical LLMs
  • Federated Learning and Critical Findings Retrieval
 
 
 
 
 
Postdoctoral Research Scholar
School of Biomedical Informatics, University of Texas Health Science Center
April 2021 – November 2023 Houston, Texas
  • Conversational Agents for Psychotherapy and Emotional Support
  • Biomedical Literature Mining and Real-time Knowledge Distillation
 
 
 
 
 
Graduate Assistant
Computer Science, University of Houston
December 2020 – August 2014 Houston, Texas
Thesis: Proactive Defense through Automated Attack Generation: A Multi-pronged Study of Generated Deceptive Content
 
 
 
 
 
Data Science-NLP Intern
Occidental (Oxy) Petroleum Corporation
May 2019 – August 2019 The Woodlands, Texas
Designed and optimized a reinforcement learning-based virtual conversational assistant to aid in digital operations and provide guided insights into system maintenance and failure detection to engineers at remote on-shore drilling rigs.
 
 
 
 
 
Summer Research Intern
Production Solutions Team, Halliburton Energy Services
May 2018 – August 2018 Houston, Texas
Implemented and designed an optimized tool for real-time automated failure and anomaly detection and prediction by leveraging supervised machine learning models trained on offshore subsea riser pipes, improving timely tool maintenance and failure detection rate by 15%.
 
 
 
 
 
Data Science Intern
2H Offshore Inc.
June 2017 – August 2017 Houston, Texas
Developed an auto-regressive neural model for fast and reliable estimation of fatigue damage in offshore drilling risers by analyzing real-time motion and temperature data from submerged sensors assessing failure estimation and maintenance.

Awards, Honors and ­Others

Cullen Graduate Success Fellowship
Hackathons
  • First Place (DecorateAR), CodeRED Discovery, University of Houston, 2018
  • Third Place , CodeRED Exploration, University of Houston, 2017
  • Winner (HarveyTrack), Social Track at HackRice 7, Rice University, 2017
Travel Grants
  • Annual Meeting of the Association for Computational Linguistics (ACL), 2020, 2022
  • Grace Hopper Conference for Women in Computing (GHC), 2015, 2016, 2018
  • International Workshop on Security and Privacy Analytics (IWSPA), 2017, 2018
  • Empirical Methods in Natural Language Processing Conference (EMNLP), 2016
  • Women in CyberSecurity Conference (WiCyS), 2016, 2017
  • Computing Research Association for Women (CRA-W), 2015
Merit-based Scholarship for Undergraduate Education

Publications

Quickly discover relevant content by filtering publications.
(2024). Exposing Vulnerabilities in Clinical LLMs Through Data Poisoning Attacks: Case Study in Breast Cancer. medRxiv.

Cite

(2024). Representation Learning of Biological Concepts: A Systematic Review. Current Bioinformatics.

PDF Cite

(2020). An in-depth benchmarking and evaluation of phishing detection research for security needs. IEEE Access.

PDF Cite

(2020). Experiments in Extractive Summarization: Integer Linear Programming, Term/Sentence Scoring, and Title-driven Models. arXiv preprint arXiv:2008.00140.

PDF Cite Custom Link