Benchmark exposes separate static and dynamic memory failures in video worlds

WorldMemBench tests whether generated worlds preserve people, objects, environments and evolving events across repeated departures and returns.

Chinese Tech
Nan Zhou · Jianxiong Gao · Qiuyu Wang · Wen Wang · Jiahao Wang · Yufeng Yuan · +7 more

Zhejiang University · Ant Group · Fudan University · Nanjing University

Research Digest··3 min read
Zhou and colleagues introduce a 260-case benchmark that evaluates memory in video world models using controlled camera departures, repeated revisits and off-screen events.

WorldMemBench contains 200 static-memory cases and 60 dynamic-memory cases across 23 semantic categories.

Why this paper

From Ant Group and 3 others

In one line

WorldMemBench evaluates video world models on static and dynamic memory retention across repeated revisits, finding no model leads all dimensions.

What we could check

  • ·No code link found
  • ·No weights link found
  • ·No dataset link found
  • ·No compute details found
  • ·No stated limitations found
  • ✓Reports numbers on named benchmarks

Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.

§

Research Digest

Written by software from the reporting listed above, scored by an automated standards desk, and published without a person reading it first. If something here is wrong, tell the editor and it will be put right.

How we workSubscribe