Madhavan Iyengar

Hi! My name's Madhavan, I'm a rising senior at the University of Michigan studying Computer Science.

I previously worked at Amazon on Alexa Smart Home, and I am currently a TA for EECS 445 (Intro to Machine Learning). In my free time you'll catch me at

Email  /  Twitter  /  LinkedIn

Research

I'm interested in building generalist embodied agents by coming up with data-efficient learning methods. I'm currently advised by Prof. Nima Fazeli and Prof. Joyce Chai.

We present an approach that uses tactile sensors to transfer policies between robots that have vastly different sensing capabilities. We show how our method can enable a cooperative task where a robot and human must work together to maneuver objects through space.

We address the lack of large-scale datasets that provide dense grounding between language and 3D scenes by introducing 3D-GRAND, a dataset comprising of 40,087 household scenes paired with 6.2 million densely-grounded scene-language instructions.

We present the first method capable of localizing novel objects in 3D scenes using Neural Radiance Field (NeRF) and Large Language Models (LLMs) through iterative, natural language-based interactions.

We address the propagation of an epidemic through a metapopulation model, highlighting the impact of migration and node degree distribution on the spread.

We explore the dynamics of disease outbreaks on college campuses, modeling interactions between extroverts and introverts. We also examine vaccination strategies, considering prioritization based on risk or contact rates to mitigate an outbreak.