Unverified50% confidenceFactExact time
The computational complexity of the full Fisher information matrix is O(p^2), where p is the number of parameters, which is infeasible for modern deep networks.
1
Sources
50%
Confidence
Long-term
Relevance
10/4/2026
First Seen
Sources
Related Entities
Related Claims
Unverified计算和存储完整Hessian矩阵的空间和时间代价是O(n²)，对数百万至数十亿参数的深度学习模型几乎不可行70% similarUnverifiedTelgarsky等人的工作证明，深度网络在表示某些函数时比等宽度的浅层网络具有指数级的参数效率优势62% similarUnverified经典计算机视觉算法的计算复杂度远低于深度神经网络，实时性强，且无需大量训练数据62% similarUnverifiedThe compound scaling mathematical formulation of EfficientNet is depth d=α^φ, width w=β^φ, and resolution r=γ^φ, where α, β, and γ are determined by grid search on the baseline model B0.61% similarVerifiedTensors, as high-dimensional generalizations of matrices, are the core data structure of deep learning frameworks such as PyTorch and TensorFlow.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/971762API
curl https://kongchang.com/api/v1/knowledge/claims/971762MCP
get_claim(id=971762)