Newest papers

Trajectory Replay: Continuous-Time Mixture Transport for Lifelong Generative Models

Alessandro De Palma, Charles Jones, Tristan Cinquin, Fabian Theis, Frederik Wenkel · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26749 · cited by 0 works (OpenAlex) · open-access PDF

German

Eine führende Strategie, generative Replay, verwendet ein generatives Modell, das auf historischen Aufgaben ausgebildet wurde, um seine eigene kontinuierliche Aktualisierung mit synthetischen Proben aus Erfahrung zu erhöhen. Es erfordert jedoch bei jedem Trainingsschritt vollständige Trajektorie-Simulationen, die Diskretisierungsfehler ansammeln und die wiedergegebene historische Verteilung degradieren können. Hier schlagen wir ein Trajektorie-Replay vor, das die Probenerzeugung umgeht, indem ein aktualisiertes Geschwindigkeitsfeld trainiert wird, um sowohl eine eingefrorene historische Geschwindigkeit entlang ihrer Trajektorien als auch das eingehende Geschwindigkeitsfeld zu entsprechen, das durch einen Stabilitäts- und Plastikkoeffizienten gewogen wird. Durch die Linsen des

Machine translation (Meta M2M100); the English original follows.

Abstract

Adapting pre-trained continuous-time generative models to sequentially incoming data without forgetting earlier tasks is a central challenge in lifelong learning. A leading strategy, generative replay, uses a generative model trained on historical tasks to augment its own continual update with synthetic samples from experience. However, it requires full trajectory simulations at every training step, which accumulate discretization errors and can degrade the replayed historical distribution. Here, we propose trajectory replay, which bypasses sample generation by training an updated velocity field to match both a frozen historical velocity along its trajectories and the incoming task's velocity field, weighted by a stability-plasticity coefficient. Through the lens of continuous-time transport, we show the resulting optimum averages the two task velocities by their respective densities and generates the mixture of their probability paths. Combined with mean velocity distillation, trajectory replay accesses historical trajectories using only a few network evaluations per continual update. Empirically, it outperforms generative replay and distillation baselines on class-incremental image generation. In two biological case studies, trajectory replay extends generative perturbation models to new contexts and experimental batches across two data modalities while preserving treatment-effect prediction.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Trajectory, Computer science, Generative grammar, Generative model, Artificial intelligence, Forgetting

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange