Bharath Muppasani

Ph.D. Candidate in Computer Science

Jan 2022 - Present
Looking for Summer Internships - 2025

Working at the intersection of Planning, Leaning and Representation at University of South Carolina, Columbia, USA.

About Me

I'm a Ph.D. candidate at the University of South Carolina, focusing on developing planning strategies for dynamic environments. My research aims to develop more efficient and adaptable planning systems that can handle complex real-world scenarios.

Resume

My research interests span across:

Automated Planning Reinforcement Learning Knowledge Representation Large Language Models

Published at: (Outlined - first authorship)

NeurIPS '24 AAAI '24 Demo Best Demo ICAPS '24 (Position, Demo, HSDIP) NeurIPS GenPlan '23 AAAI/IAAI '23 AAAI Magazine '23 IJCAI '23 Demo ICAPS PLATO '23

Featured Research

2023 - 2024

Planning Strategies for Dynamic Opinion Networks

Developed intervention strategies using ranking algorithms and neural network classifiers for accurate information dissemination in dynamic networks. Created a reinforcement learning framework analyzing multiple reward structures for network dynamics.

  • Deep Learning
  • Reinforcement Learning
  • Network Analysis
2022 - 2023

Automated Planning with Large Language Models

Spearheaded the development of datasets for fine-tuning LLMs in automated planning scenarios. Evaluated various LLM architectures for planning tasks and developed the Plansformer model.

  • PDDL
  • Large Language Models
  • Python

Selected Publications

Experience

2022 - Present

Research Assistant
University of South Carolina

Conducting research in automated planning, deep learning, and reinforcement learning. Developing novel approaches for dynamic opinion networks and planning with large language models.

  • Automated Planning
  • Machine Learning
  • Ontology
Summer 2018

Research Intern
Indian Institute of Technology, Kharagpur - Samsung Research

Conducted analysis on multi-resident smart home data, focusing on time, sensor, and frequency-based segmentation. Developed unsupervised algorithms for activity classification.

  • Data Analysis
  • Sensor Data
  • Smart Home