[KongchangAI]
Unverified50% confidenceFactExact time

In GRAC, the retrieval library consists of "graph-description" pairs from the training data, and the retrieval results are corrected by contrastive features before being used to generate descriptions of the current surgical scene.

1
Sources
50%
Confidence
Long-term
Relevance
11/9/2026
First Seen

Sources

Related Entities

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/1150309
API
curl https://kongchang.com/api/v1/knowledge/claims/1150309
MCP
get_claim(id=1150309)