Shubhankar P. Patankar

Profile.jpeg

Hello! I am a Ph.D. candidate at the intersection of cognitive science and AI with Dani S. Bassett at the University of Pennsylvania, where I also earned a Master’s degree in robotics. Before Penn, I studied mechanical engineering and Russian at the University of California, Davis.

In my research, I am interested in translating insights from human behavior to the design of autonomous learning agents capable of lifelong and open-ended learning. Currently, I am working on generating synthetic training data for offline and offline-to-online robot learning in explicitly modular environments.

Some other projects I’ve worked on in the past include:

  • Adapting theories of human curiosity as reinforcement learning rewards for agents exploring graph-structured environments [1],
  • Evaluating how humans exhibit curiosity when exploring knowledge graphs [2], and
  • Understanding the controllability properties of structural brain networks [3].

More recently, I was a machine learning intern at Tesla, working on time-series and vision problems for the 4680 cell. I’ve also been an intern at General Motors, applying graph neural networks for predictive maintenance and anomaly detection problems in vehicle manufacturing.

Outside of science and engineering, some of my interests include history, playing Age of Empires II, and supporting Liverpool Football Club.


Selected Publications

2024

  1. Mechanical prions: Self-assembling microstructures
    Mathieu Ouellet, Dani S. Bassett, Lee C. Bassett, Kieran A. Murphy, and Shubhankar P. Patankar
    arXiv, 2024
  2. Architectural styles of curiosity in global Wikipedia mobile app readership
    Dale Zhou, Shubhankar P. Patankar, David M. Lydon-Staley, Perry Zurn, Martin Gerlach, and Danielle S. Bassett
    Science Advances, 2024

2023

  1. Intrinsically motivated graph exploration using network theories of human curiosity
    Shubhankar P. Patankar, Mathieu Ouellet, Juan Cervino, Alejandro Ribeiro, Kieran A. Murphy, and Danielle Bassett
    The Second Learning on Graphs Conference, Nov 2023