Unverified50% confidenceFactExact time
At inference time, PUN employs test-time scaling: it uses the predicted global field to guide a bounded local search, thereby improving planning quality while maintaining computational efficiency.
1
Sources
50%
Confidence
Long-term
Relevance
10/21/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPUCT 在每次选择节点时，除了利用 UCT 的探索-利用权衡项，还引入了一个基于策略网络的先验预测项，用来指导搜索往更有希望的分支扩展。74% similarUnverified每套策略都经过前向滚动测试（walk-forward test）：只用历史数据挑选参数，然后在从未见过的时间段上评分，并且把手续费计算在内。70% similarUnverifiedWhen selecting tests on demand, you can use dependency-tracking tools such as Bazel or Nix, match file paths to tests with regular expressions, or have a decision model (such as Jev) predict the relevance of each test based on the diff and skip low-relevance tests after setting a probability threshold.69% similarUnverified基准测试本身就是一种数据激活的手段，为领域数据赋予清晰的优化方向69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1079183API
curl https://kongchang.com/api/v1/knowledge/claims/1079183MCP
get_claim(id=1079183)