mlsanityData to 7 Oct 2026

SD2AIL: Adversarial Imitation Learning from Synthetic Demonstrations via Diffusion Models

arXiv 2512.18583 ↗First posted 21 Dec 2025cs.LG cs.ROcited by 0

On the map: Diffusion models → Diffusion and flow policies for robotics

also in 1 more
Abstract (arXiv)

Adversarial Imitation Learning (AIL) is a dominant framework in imitation learning that infers rewards from expert demonstrations to guide policy optimization. Although providing more expert demonstrations typically leads to improved performance and greater stability, collecting such demonstrations can be challenging in certain scenarios. Inspired by the success of diffusion models in data generation, we propose SD2AIL, which utilizes synthetic demonstrations via diffusion models. We first employ a diffusion model in the discriminator to generate synthetic demonstrations as pseudo-expert data that augment the expert demonstrations. To selectively replay the most valuable demonstrations from the large pool of (pseudo-) expert demonstrations, we further introduce a prioritized expert demonstration replay strategy (PEDR). The experimental results on simulation tasks demonstrate the effectiveness and robustness of our method. In particular, in the Hopper task, our method achieves an average return of 3441, surpassing the state-of-the-art method by 89. Our code will be available at https://github.com/positron-lpc/SD2AIL.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Diffusion and flow policies for robotics, one of 14 directions in Diffusion models. Builds on 19 of its 31 references that are in the map. First posted on arXiv 21 Dec 2025.

Builds on

19 of 31 references.

Closest ancestors · 5

Same direction, newest first

  1. 2024Diffusing States and Matching Scores: A New Framework for Imitation Learning17 Oct
  2. 2024Diffusion-Reward Adversarial Imitation Learning25 May
  3. 2023DiffAIL: Diffusion Adversarial Imitation Learning11 Dec
  4. 2023Goal-Conditioned Imitation Learning using Score-based Diffusion Policies5 Apr
  5. 2022Diffusion Policies as an Expressive Policy Class for Offline Reinforcement Learning12 Aug

Other references · 14

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  3. 2016Generative Adversarial Imitation Learningcited by 1,576
  4. 2017Hindsight Experience Replaycited by 1,096
  5. 2015Prioritized Experience Replaycited by 1,070
  6. 2022Planning with Diffusion for Flexible Behavior Synthesiscited by 1,023
  7. 2017Learning Robust Rewards with Adversarial Inverse Reinforcement Learningcited by 485
  8. 2018Discriminator-Actor-Critic: Addressing Sample Inefficiency and Reward Bias in Adversarial Imitation Learningcited by 203
  9. 2019Imitation Learning via Off-Policy Distribution Matchingcited by 182
  10. 2020Primal Wasserstein Imitation Learningcited by 120
  11. 2024DiffStitch: Boosting Offline Reinforcement Learning with Diffusion-based Trajectory Stitchingcited by 35
  12. 2023A Coupled Flow Approach to Imitation Learningcited by 12
2 more
  1. 2020Robust Behavioral Cloning for Autonomous Vehicles using End-to-End Imitation Learningcited by 11
  2. 2021Hindsight Goal Ranking on Replay Buffer for Sparse Reward Environmentcited by 7

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2019Adversarial Imitation Learning from Incomplete Demonstrations
  2. 2020ADAIL: Adaptive Adversarial Imitation Learning
  3. 2022Auto-Encoding Adversarial Imitation Learning
  4. 2021Deterministic and Discriminative Imitation (D2-Imitation): Revisiting Adversarial Imitation for Sample Efficiency
  5. 2026Adversarial Imitation Learning with General Function Approximation: Theoretical Analysis and Practical Algorithms
  6. 2023Sample-efficient Adversarial Imitation Learning
  7. 2018Sample-Efficient Imitation Learning via Generative Adversarial Nets
  8. 2023Provably Efficient Adversarial Imitation Learning with Unknown Transitions