Skip to content
1762-Trajectory Replay for Continuous-Time Generative Models episode artwork

Nov 12, 2026 20 min

1762-Trajectory Replay for Continuous-Time Generative Models

from Paper Talk

The paper introduce trajectory replay, an innovative framework designed to update continuous-time generative models with sequentially arriving data without suffering from catastrophic forgetting. By replacing expensive sample-level generative replay with velocity-field regression along the historical model's own trajectories, the method halves the expected simulation cost and enhances robustness against discretization errors. Mathematically, the authors demonstrate that the optimum of this approach serves as a density-weighted mixture of historical and incoming probability paths, which effectively recovers a weighted forward KL divergence. Furthermore, the integration of a matching rule for the mean velocity enables efficient few-step sampling across the flow map. Extensive evaluations validate that trajectory replay outperforms traditional replay and distillation baselines on class-incremental image generation tasks. Finally, empirical assessments on biological case studies confirm its efficacy in extending generative perturbation models to new experimental contexts and modalities while successfully preserving treatment-effect predictions. References: Palma A, Jones C, Cinquin T, et al. Trajectory Replay: Continuous-Time Mixture Transport for Lifelong Generative Models[J]. arXiv preprint arXiv:2610.26749, 2026. 前往小宇宙评论区与主播互动

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

1762-Trajectory Replay for Continuous-Time Generative Models

0:00 20:15

No transcript yet.

Requests help us decide what to transcribe next.

Frequently asked questions

How long is this episode of Paper Talk?

This episode is 20 minutes long.

When was this Paper Talk episode published?

This episode was published on November 12, 2026.

Can I download this Paper Talk episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.