The authors synthesized training tasks that require intermediate visual evidence to determine subsequent search steps.
Neos turns visual evidence into better follow-up search decisions
The framework jointly trains a search policy and visual evidence interface to support long, multimodal research tasks.
Chinese Tech
Ekko He · Jack Hong · Shilin Yan · Xintong Zhang · Yujie Zhang · Wenjie Li · +5 more
Fudan University · Accio Team, Alibaba Group · Shanghai Innovation Institute
Research Digest··2 min read
He and colleagues present Neos, a multimodal search framework designed to make retrieved images actively shape later queries and tool use.
Why this paper
From Accio Team, Alibaba Group and 2 others
In one line
Neos learns to interleave search and visual reasoning, achieving leading performance on multimodal deep search benchmarks.
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.
§