Unverified50% confidenceFactExact time
The cumulative effective learning rate C_T = Σ_{t<T} η_t / ||W_t||_F links weight-norm control with the optimizer's exploration capability.
1
Sources
50%
Confidence
Long-term
Relevance
10/28/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLISCO (Learned Iterative Solver for Constrained Optimization) learns to predict primal-dual step sizes to iteratively approach a solution satisfying KKT optimality conditions, rather than directly predicting the final control output.66% similarUnverifiedThe online estimation algorithm studied in the paper is a single-pass stochastic gradient algorithm with a constant learning rate γ, in the recursive form θ_t = θ_{t-1} + γ * ψ(Y_t, θ_{t-1}), where ψ is the expectile score function.65% similarUnverified重参数化技巧将采样表示为 z = μ + σ × ε（ε从标准正态分布采样），使随机性外置于ε，从而让μ和σ的梯度可正常反向传播实现端到端训练64% similarUnverifiedHimanshu指出LLM推理管线的最大优势在于可微分，可用强化学习优化，但无法学习代数构造64% similarUnverifiedThe study generates samples of X_{t+Δt} conditioned on X_t, allowing the network to implicitly learn the conditional distribution p(X_{t+Δt}|X_t) by training on observed segments64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1106005API
curl https://kongchang.com/api/v1/knowledge/claims/1106005MCP
get_claim(id=1106005)