Postdoctoral scholar, reinforcement learning, partial observability, causality
gym-gridverse. Gridworld domains in the gym interface
29beamerthemeNU. Beamer theme for Northeastern University
16asym-rlpo. Asymmetric methods for partially observable reinforcement learning
10gym-pomdps. Python
10rl-rpsr. Code for reward-predictive state representations
2rl-parsers. Python
2pyfgraph. Generic factor graph implementation
2discovery-utils. utilities to run and look at experiments on discovery
1baise.ro. My personal website, written with python/Flask
1gym-pyro. OpenAI Gym environments for MDPs, POMDPs, and confounded-MDPs implemented as pyro-ppl probabilistic programs.
1pymarl2. Fine-tuned MARL algorithms on SMAC (100% win rates on most scenarios)
1one-to-one. Bijections between semantic value and index
1fittk. CLI toolkit for inspecting and correcting Garmin swim FIT files: fix mislabelled lengths, drop spurious laps, and recompute all aggregates.
1