Maximum Entropy Gain Exploration for Long Horizon Multi-goal Reinforcement Learning
We propose an objective for multi-goal reinforcement learning to encourage the agents to pursue sparsely explored past achieved goals to focus the exploratio...
Presentations
2020
2018
Curiosity-driven Exploration by Self-supervised Prediction
A paper review presentation for CSC2541 Deep RL
GraphRNN: A Deep Generative Model for Graphs
A paper review presentation for CSC2547. Co-present with Jesse Bettencourt.
In-Place Activated BatchNorm for Memory-Optimized Training of DNNs
A paper review presentation for CSC2548