Mixture of Contexts for Long Video Generation
This work addresses the challenge of generating coherent long videos for applications in media and AI, representing a novel method for a known bottleneck rather than a foundational breakthrough.
The paper tackles the problem of long video generation by addressing the quadratic cost of self-attention in diffusion transformers, proposing Mixture of Contexts (MoC) as a learnable sparse attention routing module to enable efficient memory retrieval and consistency over minutes of content.
Long video generation is fundamentally a long context memory problem: models must retain and retrieve salient events across a long range without collapsing or drifting. However, scaling diffusion transformers to generate long-context videos is fundamentally limited by the quadratic cost of self-attention, which makes memory and computation intractable and difficult to optimize for long sequences. We recast long-context video generation as an internal information retrieval task and propose a simple, learnable sparse attention routing module, Mixture of Contexts (MoC), as an effective long-term memory retrieval engine. In MoC, each query dynamically selects a few informative chunks plus mandatory anchors (caption, local windows) to attend to, with causal routing that prevents loop closures. As we scale the data and gradually sparsify the routing, the model allocates compute to salient history, preserving identities, actions, and scenes over minutes of content. Efficiency follows as a byproduct of retrieval (near-linear scaling), which enables practical training and synthesis, and the emergence of memory and consistency at the scale of minutes.