Entity · paper

The Seriality Gap in Video Diffusion Models

paperactiveprovisionalthe-seriality-gap-in-video-diffusion-models-00b3e7a1·1 events·first seen 16h ago

Aliases: The Seriality Gap in Video Diffusion Models

More like this (12)

Masked Diffusion Models discrete diffusion models Diffusion Models Survival Diffusion Probabilistic Model (SDPM)What Does a Discrete Diffusion Model Learn?Representation-Conditioned Diffusion Models latent diffusion model Self-Augmenting Retrieval for Diffusion Language Models Diffusion Language Models A Diffusion Approximation for Temporal-Difference Learning with Linear Features under Markovian Noise Denoising Diffusion Probabilistic Models text-to-image diffusion model

Recent events (1)

6arXiv · cs.LG·16h ago·source ↗

Seriality gap identified as structural limitation of video diffusion models on causal chain tasks

A new arXiv paper identifies and formalizes the 'seriality gap' in video diffusion models: a structural mismatch where tasks requiring growing serial computation (e.g., multi-ball physics chains) degrade as causal chain length increases, even with more denoising steps. Controlled experiments on hard-sphere dynamics isolate dependent-event structure—not video length—as the cause. The authors prove that for deterministic video prediction, denoising steps add no serial computation beyond the backbone, and show that autoregressive/blockwise generation and architectural depth help mitigate the gap. This has direct implications for using video diffusion as a world model or physics simulator.

Evaluation and Benchmarking Multimodal Progress The Seriality Gap in Video Diffusion Models