Zero-Shot Transfer of Haptics-Based Object Insertion Policies

1 code implementation • 29 Jan 2023 • Samarth Brahmbhatt, Ankur Deka, Andrew Spielberg, Matthias Müller

In this paper we train a contact-exploiting manipulation policy in simulation for the contact-rich household task of loading plates into a slotted holder, which transfers without any fine-tuning to the real robot.

ARC -- Actor Residual Critic for Adversarial Imitation Learning

no code implementations • 5 Jun 2022 • Ankur Deka, Changliu Liu, Katia Sycara

In AIL, an artificial adversary's misclassification is used as a reward signal that is optimized by any standard Reinforcement Learning (RL) algorithm.

Continuous Control Imitation Learning +1

Natural Emergence of Heterogeneous Strategies in Artificially Intelligent Competitive Teams

1 code implementation • 6 Jul 2020 • Ankur Deka, Katia Sycara

We corroborate that modeling agents with Graph Neural Networks and training them with Reinforcement Learning leads to the evolution of increasingly complex strategies for each team.