Unverified50% confidenceFactExact time
SSM用固定维度d的隐状态压缩长度为N的历史序列时，根据率失真理论存在信息保留能力的理论上限，当N远大于d时必然产生信息丢失
1
Sources
50%
Confidence
Long-term
Relevance
8/12/2026
First Seen
Sources
Related Claims
Unverified根据率失真理论，若序列包含O(N)比特相关信息而隐状态只有O(d)比特容量，当N远大于d时必然产生信息丢失76% similarUnverified精确的 K 近邻搜索（KNN）时间复杂度为 O(n×d)，在数百万级别记忆条目下延迟不可接受，ANN 算法通过牺牲极小精度换取数量级速度提升68% similarUnverifiedThe choice of bottleneck dimensionality is essentially a trade-off between information compression rate and reconstruction quality, corresponding to rate-distortion theory; Shannon proved in 1959 that there exists a theoretical minimum coding rate lower bound for a given distortion tolerance.67% similarUnverified只有在LLM推理比embedding加向量检索慢数十倍以上时，语义缓存才能同时实现降本与提速67% similarUnverifiedRNN存在两大缺陷：无法并行计算导致速度慢，以及长距离依赖信息衰减的健忘问题67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/734755API
curl https://kongchang.com/api/v1/knowledge/claims/734755MCP
get_claim(id=734755)