The authors review memory mechanisms for models that generate videos sequentially as tokens, frames, or chunks.
Persistent memory is essential for coherent autoregressive video generation
The authors organize existing memory mechanisms by representation, purpose, operation, training, and evaluation, while identifying requirements for reliable long-term generation.
Big Tech
Harold Haodong Chen · Rongjin Guo · Disen Lan · Wen-Jie Shu · Hongfei Zhang · Hanzhe Hu · +19 more
HKUST · CityUHK · FDU · ZODA · CMU
Research Digest··2 min read
Chen and colleagues survey how autoregressive video generators can preserve relevant history after it leaves their limited context windows.
Why this paper
From NVIDIA and 15 others
In one line
Memory in autoregressive video generation is a five-part problem: forms, functions, operations, learning, and evaluation.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§