Research map / Reinforcement learning / Imitation learning

PN-GAIL: Leveraging Non-optimal Information from Imperfect Demonstrations

ICLR 2025 · Reinforcement learning › Imitation learning · approach: Offline imitation from imperfect data

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes PN-GAIL, a generative adversarial imitation learning method that leverages both optimal and non-optimal demonstrations from imperfect data.

Problem

Imitation learning typically assumes optimal expert demonstrations, but real-world demonstration data contains a mix of optimal and non-optimal examples.

Approach

Extends adversarial imitation learning to exploit non-optimal demonstrations rather than discarding them, using their information alongside optimal demonstrations to learn a policy.

Experimental setup

Builds on: GAIL

Closest papers by meaning

• On Computation and Generalization of Generative Adversarial Imitation Learning (ICLR 2020 · similarity 0.93)
• Exploring Gradient Explosion in Generative Adversarial Imitation Learning: A Probabilistic Perspective (AAAI 2024 · similarity 0.93)
• Generative Adversarial Imitation Learning with Neural Network Parameterization: Global Optimality and Convergence Rate (ICML 2020 · similarity 0.92)
• C-GAIL: Stabilizing Generative Adversarial Imitation Learning with Control Theory (NeurIPS 2024 · similarity 0.91)
• Diffusion-Reward Adversarial Imitation Learning (NeurIPS 2024 · similarity 0.89)
• Provably and Practically Efficient Adversarial Imitation Learning with General Function Approximation (NeurIPS 2024 · similarity 0.88)