[KongchangAI]
Unverified50% confidenceBenchmarkExact time

On the temporal split test set (878 comparisons from 655 trials), a gradient-boosted emulator trained on registry data eliminated about one-fifth of the error of a simple baseline.

1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/5/2026
First Seen
Valid until: 2/3/2027

Sources

Related Entities

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/1137966
API
curl https://kongchang.com/api/v1/knowledge/claims/1137966
MCP
get_claim(id=1137966)