Unverified50% confidenceFactExact time
To improve the stability of the Hebbian rule, researchers proposed the Oja rule (introducing weight decay to achieve normalization) and the BCM rule (introducing a sliding modification threshold).
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPure Hebbian rules face an instability problem: positive feedback may cause weights to grow without bound.65% similarUnverifiedThe study adopted a confidence-interval-aware qualification rule, meaning a configuration passes only when its lower confidence bound genuinely exceeds the threshold τ.56% similarUnverifiedThe study proposes a recalibration protocol based on a precision-floor approach to correct threshold bias.54% similarUnverifiedOOSM 除回溯重放外还有 Algorithm A（用当前增益近似处理）和 Algorithm B（基于储存协方差进行精确后验修正）等方案，各有计算复杂度与精度权衡53% similarUnverifiedThe strategy of rate-limiting rather than banning reflects the principle of proportionality, with the penalty ladder being: label reminder → reduced recommendation weight → restricted interaction features → temporary ban → permanent ban.53% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1021836API
curl https://kongchang.com/api/v1/knowledge/claims/1021836MCP
get_claim(id=1021836)