Unverified50% confidenceFactExact time
The decomposition strategy of spatiotemporal attention reduces computational complexity from O((T×H×W)²) to O(T×(H×W)²+(H×W)×T²).
1
Sources
50%
Confidence
Long-term
Relevance
9/10/2026
First Seen
Sources
Related Entities
Related Claims
Unverified滑动窗口注意力由 Beltagy 等人在 Longformer（2020）中提出，将复杂度从 O(n²) 降至 O(n×w)74% similarUnverifiedA domain decomposition strategy can reduce the step complexity from O((L/w)²) to O(L²/(N·w²)), where N is the number of subdomains.71% similarVerifiedThe self-attention mechanism causes computational complexity to grow quadratically with sequence length, so expanding the context from 1,000 to 4,000 tokens increases the computation by 16 times.69% similarUnverifiedThe root cause of attention dilution lies in Softmax normalization: the more tokens in a sequence, the effective attention weight allocated to each position decreases approximately inversely proportional to the sequence length.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/889210API
curl https://kongchang.com/api/v1/knowledge/claims/889210MCP
get_claim(id=889210)