Unverified50% confidenceFactExact time
Deep Blue could evaluate approximately 200 million chess positions per second, employed an Alpha-Beta pruning search algorithm, and was equipped with 480 dedicated chips with hardware-level optimization for chess evaluation functions.
1
Sources
50%
Confidence
Long-term
Relevance
8/31/2026
First Seen
Sources
Related Entities
Related Claims
Unverified建议建立人工标注的黄金基准集通常包含 200-500 张精心标注的样本，用于持续监控自动标注管线的性能漂移63% similarUnverified该项目对每个可达棋盘状态调用Minimax求解器计算最佳落子，形成棋盘状态到最佳落子的映射对作为神经网络监督学习标签62% similarUnverifiedObservability tools such as Helicone, LangSmith, Portkey, and Braintrust help enterprises track the cost, latency, and quality of each LLM call.58% similarUnverified质心追踪算法通过计算当前帧质心与上一帧已注册目标质心的欧氏距离进行配对，计算复杂度随目标数量呈线性甚至平方级增长58% similarUnverifiedCodex使用了126个子智能体、32500次工具调用、391次浏览器测试和2300多次单元测试；Claude Code使用35个子智能体、102次浏览器测试和296次单元测试58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/829650API
curl https://kongchang.com/api/v1/knowledge/claims/829650MCP
get_claim(id=829650)