PulseAugur
实时 08:24:33
English(EN) The Price of Hidden Curvature: An $\widetildeΩ (d^{5/4} \sqrt{T})$ Lower Bound for Bandit Convex Optimization

新下界表明 Bandit 凸优化比线性 Bandit 更难

研究人员为 Bandit 凸优化建立了一个新的下界,证明它比线性 Bandit 的复杂性根本上更高。该研究引入了一类新的凸函数,揭示了一种权衡:学习者要么在不知道关键线性变换的情况下发现特定的“管”,要么花费观察来学习它。这导致找到 $\varepsilon$-最优动作的样本复杂度下界为 $\widetilde{\Omega}(d^{5/2}/\varepsilon^2)$,转化为遗憾下界为 $\widetilde{\Omega}(d^{5/4}\sqrt{T})$。这些发现扩展到无约束动作空间。 AI

影响 为某些优化问题的学习效率设定了理论上限,可能指导未来的算法开发。

排序理由 该集群包含一篇详细介绍具有新数学界限的理论研究的学术论文。

在 Hugging Face Daily Papers 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

新下界表明 Bandit 凸优化比线性 Bandit 更难

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
该集群包含一篇详细介绍具有新数学界限的理论研究的学术论文。
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
paper, other
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
48 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准

报道来源 [2]

  1. arXiv cs.LG TIER_1 English(EN) · Nived Rajaraman ·

    隐藏曲率的代价:Bandit凸优化中的 $\widetilde{\Omega} (d^{5/4} \sqrt{T})$ 下界

    arXiv:2607.18652v1 Announce Type: cross Abstract: We establish a $\widetilde\Omega(d^{5/4}\sqrt T)$ lower bound on the minimax expected regret of stochastic bandit convex optimization of $1$-Lipschitz functions on the Euclidean ball. This presents the first nontrivial regret lowe…

  2. Hugging Face Daily Papers TIER_1 English(EN) ·

    隐藏曲率的代价:用于Bandit凸优化的 $\widetildeΩ (d^{5/4} \sqrt{T})$ 下界

    We establish a $\widetildeΩ(d^{5/4}\sqrt T)$ lower bound on the minimax expected regret of stochastic bandit convex optimization of $1$-Lipschitz functions on the Euclidean ball. This presents the first nontrivial regret lower bound that grows faster than $d\sqrt{T}$ for this pro…