Search this site
Embedded Files
Dr. Tirthankar Ghosal (তীর্থঙ্কর ঘোষাল)
  • Home
  • Research and Projects
    • AI for Science
    • AI for Operations
  • Publications
  • Community Activities
  • Teaching
  • Achievements
  • Curriculum Vitae
Dr. Tirthankar Ghosal (তীর্থঙ্কর ঘোষাল)
  • Home
  • Research and Projects
    • AI for Science
    • AI for Operations
  • Publications
  • Community Activities
  • Teaching
  • Achievements
  • Curriculum Vitae
  • More
    • Home
    • Research and Projects
      • AI for Science
      • AI for Operations
    • Publications
    • Community Activities
    • Teaching
    • Achievements
    • Curriculum Vitae

AI for Science

Scientific Hypothesis Generation (Material Science)

Team Members: Emily Heron (ORNL), Miaosen Chai (USC), Tom Hope (Al2) [PI], Tirthankar Ghosal [Co-PI]

Training LLMs to extrapolate to generate "new hypotheses". Reinforcement Learning with Verifiable Rewards (RLVR).

AstroMLab

Principal Team Members: Yuan-Sen Ting (OSU) [PI], Tirthankar Ghosal [Co-PI], Tijmen de Haan (KEK-Japan) [Co-PI], Alberto Aaccomazzi (Harvard CfA), Emily Herron (ORNL), Vanessa Lama (ORNL), Josh Nguyen (UChicago), Rui Pan (UIUC), Hardik Arora (IIT Patna)

A consortium of Astrophysicists, Computer Scientists, Data Archivists, Astronomers from Ohio State University, Oak Ridge National Laboratory, NASA Astrophysics Data System, KEK-Japan, Argonne National Laboratory, working together on several problems of Astrophysics and Astronomy, where ML plays a central role. https://astromlab.org/ 

Awarded several Director's Discretionary (DD) compute allocation awards at ORNL.

  • AST 209

  • AST 216

  • AST 221

  • AST 228

The collaboration resulted in the development of Astrosage, AstroLLaMa models, AstroBench. More details soon!

mCLM v2.0: A Function-Infused and Synthesis-Friendly Modular Chemical Language Model

Collaborators: Heng Ji (UIUC) [PI]


Agentic AI for Scientific Discovery (Additive Manufacturing)

Collaborators: Stephen DeWitt (ORNL), Daniel Rosendo (ORNL), Renan De Souza (ORNL), Philippe Austria (ORNL), Sarp Oral (ORNL), Ben Mintz (ORNL), Rafael Fernando (ORNL)

Demonstration of Agentic AI capabilities at MDF, ORNL with 3D Printers. This is a multi-division collaboration within ORNL to explore AI agents in the closed loop scientific workflows with a Additive Manufacturing use-case.

AI-Driven Autonomous Exploration of Dark Matter and Dark Energy Through Cosmological Simulations

Collaborators: Brant Robertson (UCSC) [PI], Yuan-Sen Ting (OSU) [Co-PI]

Application for DOE INCITE 2026 allocations. Recently awarded a ORNL-DD compute allocation. To start from September 2025.

Accelerating Scientific Discovery Through Z-score-informed Multi-agent Large Language Models for Scientific Hypothesis Generation

Collaborators: Jian Wu (Old Dominion University) [PI], Sarah Rajtmajer (PSU), Dominik Soos (ODU)

Awarded a DD compute grant at ORNL. To start from September 2025.

Iterative Construction of an Automated Synthesis Knowledge Graph

Collaborators: Elsa Olivetti (MIT) [PI], Vineeth Venugopal (MIT) [Co-PI]

Awarded a DD compute grant at ORNL: MAT721

EPIC: Generative AI Platform for Accelerating HPC Operational Data Analytics

Collaborators: Ahmad Maroof Karimi (ORNL), Woong Shin (ORNL), Jesse Hines (ORNL)

Develop a GenAI based operational data analytics tool for leadership class supercomputers.

Towards AI-Powered Autonomous Labs for Neutron Science and Quantum Materials

Team: Tirthankar Ghosal [PI], Yongqiang Chen [Co-PI] (ORNL)

Application for Seed Proposal at ORNL (under development)

Retrieval Augmented Generation for Manufacturing and Manufacturing Assistant for Augmented Reality Glasses

Collaborators: Pum Kim (ORNL) [Co-PI], Komal Chawla (ORNL) [PI]

Application for Seed Proposal at ORNL (under development)

Generative AI for Omics (Spatial Transcriptomics and Proteomics)

Collaborators: Pratik Dutta (Stony Brook University)


Universe TBD

A consortium of ML researchers, Astronomers, and Computer Scientists 👉 UniverseTBD 

  • Hypothesis Generation

  • LLM Deception

  • JWST Bench

LinkLinkedInLinkLinkLink

☎️+1(865) 341-2922

📧 <last-name>t at ornl dot gov

📧t<last-name> at acm dot org

📧t<last-name> at utk dot edu

Google Sites
Report abuse
Page details
Page updated
Google Sites
Report abuse