Unverified50% confidenceFactExact time
Deep Blue relies mainly on brute-force search, exhaustively computing optimal solutions within a limited depth via dedicated chips, and is essentially rule-driven computation.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified深蓝采用极小化极大算法（Minimax）配合Alpha-Beta剪枝作为核心技术65% similarUnverifiedDeepMind的AlphaCode和AutoML体系使用进化搜索确定网络拓扑，再用梯度下降训练权重61% similarUnverifiedIBM's Deep Blue defeated Kasparov in 1997 and made extensive use of expert knowledge such as opening libraries and evaluation functions from human chess players.61% similarUnverifiedIn scenarios with scarce data and limited compute, the hybrid architecture of search plus learning has a significant advantage over end-to-end deep learning in terms of sample efficiency.59% similarUnverifiedThe core idea of the deep learning lane detection algorithm introduced in the article is to simplify dense segmentation into efficient classification.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/908058API
curl https://kongchang.com/api/v1/knowledge/claims/908058MCP
get_claim(id=908058)