On Training Flexible Robots using Deep Reinforcement Learning
The use of robotics in controlled environments has flourished over the last several decades and training robots to perform tasks using control strategies developed from…
Publications
Goal-conditioned Imitation Learning
Designing rewards for Reinforcement Learning (RL) is challenging because it needs to convey the desired task, be efficient to optimize, and be easy to compute.…
Channel Attention Networks
In this work, we propose Channel Attention Networks (CAN), a deep learning model that uses soft attention on individual channels. We jointly train this model…
Using A Priori Knowledge to Improve Scene Understanding
Existing semantic segmentation algorithms treat images in isolation, but autonomous vehicles often revisit the same locations. We propose leveraging this a priori knowledge to improve…
Hierarchical Policy Learning Is Sensitive To Goal Space Design
Hierarchy in reinforcement learning agents allows for control at multiple time scales yielding improved sample efficiency, the ability to deal with long time horizons and…
Label Efficient Audio Classification Through Multitask Learning And Self Supervision
While deep learning has been incredibly successful in modeling tasks with large, carefully curated labeled datasets, its application to problems with limited labeled data remains…
Heuristics For Image Generation From Scene Graphs
Generating realistic images from scene graphs requires neural networks to be able to reason about object relationships and compositionality. Learning a sufficiently rich representation to…
Graph-DQN: Fast Generalization To Novel Objects Using Prior Relational Knowledge
Humans have a remarkable ability to both generalize known actions to novel objects, and reason about novel objects once their relationship to known objects is…
SPIGAN: Privileged Adversarial Learning from Simulation
Deep Learning for Computer Vision depends mainly on the source of supervision. Photo-realistic simulators can generate large-scale automatically labeled synthetic data, but introduce a domain…