H
Homanga Bharadhwaj
Researcher at University of Toronto
Publications - 52
Citations - 587
Homanga Bharadhwaj is an academic researcher from University of Toronto. The author has contributed to research in topics: Reinforcement learning & Computer science. The author has an hindex of 11, co-authored 45 publications receiving 285 citations. Previous affiliations of Homanga Bharadhwaj include Indian Institute of Technology Kanpur & Carnegie Mellon University.
Papers
More filters
Proceedings ArticleDOI
Meta-Learning for User Cold-Start Recommendation
TL;DR: This work designs a recommendation framework that is trained to be reasonably good enough for a wide range of users and handles the user cold-start model much better than state-of-the art benchmark recommender systems.
Posted Content
Conservative Safety Critics for Exploration
Homanga Bharadhwaj,Aviral Kumar,Nicholas Rhinehart,Sergey Levine,Florian Shkurti,Animesh Garg +5 more
TL;DR: This paper theoretically characterize the tradeoff between safety and policy improvement, show that the safety constraints are likely to be satisfied with high probability during training, derive provable convergence guarantees for the approach, and demonstrate the efficacy of the proposed approach on a suite of challenging navigation, manipulation, and locomotion tasks.
Proceedings ArticleDOI
RecGAN: recurrent generative adversarial networks for recommendation systems
TL;DR: This work uses customized Gated Recurrent Unit cells to capture latent features of users and items observable from short-term and long-term temporal profiles and proposes a Recurrent Generative Adversarial Network (RecGAN), which outperforms other baseline models irrespective of user behavior and density of training data.
Posted Content
DiVA: Diverse Visual Feature Aggregation for Deep Metric Learning
Timo Milbich,Karsten Roth,Karsten Roth,Homanga Bharadhwaj,Homanga Bharadhwaj,Samarth Sinha,Samarth Sinha,Yoshua Bengio,Yoshua Bengio,Björn Ommer,Joseph Paul Cohen +10 more
TL;DR: This work proposes and studies multiple complementary learning tasks, targeting conceptually different data relationships by only resorting to the available training samples and labels of a standard DML setting, resulting in strong generalization and state-of-the-art performance on multiple established DML benchmark datasets.
Proceedings ArticleDOI
A Data-Efficient Framework for Training and Sim-to-Real Transfer of Navigation Policies
TL;DR: This work introduces a robust framework that plans in simulation and transfers well to the real environment, consisting of the encoder and planner modules, and shows successful planning performances in different navigation tasks.