DecMem: Towards Minute-Long Consistent World Generation with Decoupled Memory
:::info Stub — Full Engineering Breakdown Coming This paper has a linked code implementation and was featured on Hugging Face Papers with 8 upvotes. A full breakdown with production viability rating, implementation notes, and honest limitations is being written. Subscribe to AI Letters → :::
| Authors | Zhenhao Yang et al. |
| Year | 2026 |
| HF Upvotes | 8 |
| arXiv | 2605.31336 |
| Download | |
| Code | https://github.com/KlingAIResearch/DecMem |
Abstract
Recent advances in video generative models have promoted rapid progress in controllable world models. However, maintaining fine-grained spatio-temporal consistency under long-horizon reasoning remains a key challenge. In this work, we move beyond explicit 3D memory and coarse frame-level implicit modeling, and propose a fine-grained, learnable, and scalable memory for consistent world generation. We first identify two fundamental limitations of naïve learnable memory architectures in long-horizon extrapolation, namely computational inefficiency and attention dispersion. Through a systematic analysis of attention dispersion, we propose DecMem, a decoupled memory architecture that employs Sparse Global Memory for efficient fine-grained access to global history and Anchored Local Memory for stable and high-quality extrapolation. Extensive experiments demonstrate that DecMem significantly outperforms current state-of-the-art methods. By ensuring precise and efficient long-term memory and achieving superior extrapolation capabilities, DecMem enables minute-level controllable long video generation with high fidelity and consistency.
Engineering Breakdown
The Problem
However, maintaining fine-grained spatio-temporal consistency under long-horizon reasoning remains a key challenge.
The Approach
In this work, we move beyond explicit 3D memory and coarse frame-level implicit modeling, and propose a fine-grained, learnable, and scalable memory for consistent world generation. Through a systematic analysis of attention dispersion, we propose DecMem, a decoupled memory architecture that employs Sparse Global Memory for efficient fine-grained access to global history and Anchored Local Memory for stable and high-quality extrapolation.
Key Results
Extensive experiments demonstrate that DecMem significantly outperforms current state-of-the-art methods.
Research Areas
This paper contributes to the following areas of AI/ML engineering:
- Machine learning
- Deep learning
- Neural networks
- Model optimization
- AI systems
- Minutelong
:::tip Subscribe Get weekly breakdowns of papers like this in AI Letters - the newsletter for engineers building production AI systems. :::