profile photo
Shubham Sharma

I was a research intern at CLVR lab at KAIST for a year advised by Prof. Joseph Lim working on various topics in robot learning. I graduated from IIT Bombay with a major in Aerospace Engineering (honour student) and a minor in AI. During my undergrad, I was advised by Prof. Xianyuan Zhan and Prof. Biplab Banerjee at IIT Bombay and worked on algorithms for offline reinforcement learning.

I also led a student team that worked on autonomous drones đŸ•šī¸đŸš!
(memories with the team)

My goal is to develop robots 🤖 with human-like intelligence which includes embodying them with congnitive capabilities for understanding reality 🧠 and physical interlligence to influence it đŸĻž. Specifically i am interested in:

  • Learning robust sensorimotor skills (dexterous manipulation ,whole-body control, visuotactile manipulation)
  • Adapting learned skills to new environments (rapid adaptation, continual learning)
  • Composing skills for more complex tasks (skill-chaining, energy-based models)

Intel

Research Intern

KAIST AI

2023-Current

Tsinghua University

Research Intern (Remote)

Tsinghua University

2022-2023

Daikin RnD Japan

Intern (Remote)

Daikin RnD Japan

Summer 2022

IIT Bombay

BTech, AE

IIT Bombay

2019 - 2023

Publications

When to Trust Your Simulator: Dynamics-Aware Hybrid Offline-and-Online Reinforcement Learning

Haoyi Niu,Shubham Sharma,Yiwen Qiu,Ming Li,Guyue Zhou,Jianming Hu, Xianyuan Zhan
NeurIPS 2022 (Spotlight)
Paper  •   Code   •   Video


A Novel Approach to Autonomous Aerial Manipulation using Multiple Drones

Team Aerove, UMIC
IARC 2021 Symposium on Autonomous aerial flight
Design Report

Projects

Automating unstructured welding

Shubham Sharma, Junseung Lee
Worked on motion planning soltuions for automating precision welding using a RB3 arm and depth data

Muscle memory: retreival augumented RL for dexterous manipulation

Minho Heo, Shubham Sharma
We try to speed up RL in high-dimentional control problems via retreival of action synergies
Report  •   Slides

International Aerial Robotics Challenge (Mission 9)

Innovation Cell IITB
Led a team of 20+ undergrads in the making of a mother-daughter drone system to attempt performing a precise insertion task mid-air
Blog with videos

Service

Reviewer for CoRL2024