4 papers · 1 filter
On the Stability and Realizability of Recurrent Polynomial Surrogate Ternary Logic Gate Networks
Sai Sandeep Damera, Ryan Matheu, Aniruddh G. Puranic +2
Recurrent Neural Networks (RNNs) can learn to predict Signal Temporal Logic (STL) verdicts online from partial trajectories, but deploying them as runtime monitors in safety-critic…
Learning from Imperfect Demonstrations via Temporal Behavior Tree-Guided Trajectory Repair
Aniruddh G. Puranic, Sebastian Schirmer, John S. Baras +1
Learning robot control policies from demonstrations is a powerful paradigm, yet real-world data is often suboptimal, noisy, or otherwise imperfect, posing significant challenges fo…
Safety-Aware Reinforcement Learning for Control via Risk-Sensitive Action-Value Iteration and Quantile Regression
Clinton Enwerem, Aniruddh G. Puranic, John S. Baras +1
Mainstream approximate action-value iteration reinforcement learning (RL) algorithms suffer from overestimation bias, leading to suboptimal policies in high-variance stochastic env…
Towards Efficient Risk-Sensitive Policy Gradient: An Iteration Complexity Analysis
Rui Liu, Anish Gupta, Erfaun Noorani +1
Reinforcement Learning (RL) has shown exceptional performance across various applications, enabling autonomous agents to learn optimal policies through interaction with their envir…