Research map / Reinforcement learning / MDPs and planning
Langevin Dynamics for Adaptive Inverse Reinforcement Learning of Stochastic Gradient Algorithms
Vikram Krishnamurthy, George Yin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use passive Langevin IRL (no control over gradient evaluation point) instead of classical Langevin dynamics | Bayesian posterior reconstruction · 2-D Bayesian learning | variational distance to classical Langevin: d(1) = 0.0122, d(2)= 0.0202 | No clear effect |
| use two-time-scale multi-kernel IRL without knowing the initialization density | passive Langevin IRL · 2-D Bayesian learning | posterior reconstruction quality | No clear effect |
| use passive IRL algorithms instead of classical Langevin on a high-dimensional posterior | passive Langevin IRL · a9a logistic regression (N = 124) | L1 distance of marginals to classical Langevin | No clear effect |
| add multiple parallel inverse learners (ablation) | passive Langevin IRL · high-dimensional problems | reconstruction quality | No clear effect |
Closest papers by meaning
- Efficient Exploration of Reward Functions in Inverse Reinforcement Learning via Bayesian Optimization (NeurIPS 2020 · similarity 0.88)
- Maximum Entropy Inverse Reinforcement Learning of Diffusion Models with Energy-Based Models (NeurIPS 2024 · similarity 0.88)
- Maximum-Likelihood Inverse Reinforcement Learning with Finite-Time Guarantees (NeurIPS 2022 · similarity 0.88)
- Provable and Practical: Efficient Exploration in Reinforcement Learning via Langevin Monte Carlo (ICLR 2024 · similarity 0.88)
- How does Inverse RL Scale to Large State Spaces? A Provably Efficient Approach (NeurIPS 2024 · similarity 0.87)
- Recent Advancements in Inverse Reinforcement Learning (AAAI 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
