PulseAugur
中
实时 21:53:05

新的 Bandit 算法解决了公平性和连续 K-Max 问题 · 跟踪 6 个来源

作者 PulseAugur 编辑部 · [10 个来源] ·

研究人员开发了新的 Bandit 问题算法，这些算法用于推荐系统等应用。对于连续 K-Max bandits，DCK-UCB 实现了 $\widetilde{O}(T^{3/4})$ 的次线性 regret 边界，而 MLE-Exp 为指数分布提供了接近最优的 $\widetilde{O}(\sqrt{T})$ regret。在相关领域，UCB-HARE 解决了 bandits 中的“公平性成本”问题，建立了 $\Omega(\sigma\sqrt{k^{\max(1,q)}/T})$ 的紧凑式 minimax 下界和一个与之匹配的算法。此外，一种新颖的 Bandit PCA 算法实现了高达对数因子（polylogarithmic factors）的 $r\sqrt{dT}$ 的 minimax regret，改进了先前的界限并与量子层析成像（quantum tomography）相关联。 AI

影响 推进了对复杂 Bandit 问题的理论理解和算法解决方案，可能改进推荐和决策系统。

排序理由 多篇 arXiv 论文详细介绍了 Bandit 问题的新算法和理论保证。

在 arXiv stat.ML 阅读 →

AI 生成摘要 · Google Gemini · 来自 10 个来源。 我们如何撰写摘要 →

新的 Bandit 算法解决了公平性和连续 K-Max 问题 · 跟踪 6 个来源

报道来源 [10]

  1. arXiv cs.LG TIER_1 English(EN) · Xin Li, Zixin Zhong ·

    MESHA：用于战略线性老虎机的机制强制顺序减半

    arXiv:2607.14706v1 Announce Type: new Abstract: We design and analyze \underline{M}echanism-\underline{E}nforced \underline{S}equential \underline{HA}lving (MESHA), an algorithm for Best Arm Identification (BAI) in strategic linear bandits. In this setting, each arm may strategic…

  2. arXiv cs.LG TIER_1 English(EN) · Zixin Zhong ·

    MESHA：用于战略性线性老虎机的机制强制顺序折半

    We design and analyze \underline{M}echanism-\underline{E}nforced \underline{S}equential \underline{HA}lving (MESHA), an algorithm for Best Arm Identification (BAI) in strategic linear bandits. In this setting, each arm may strategically misreport its feature vector to maximize th…

  3. arXiv cs.LG TIER_1 English(EN) · Yu Chen, Siwei Wang, Longbo Huang, Wei Chen ·

    连续K-最大老虎机问题的次线性遗憾

    arXiv:2502.13467v2 Announce Type: replace Abstract: The $K$-Max combinatorial multi-armed bandit problem arises in applications such as recommendation and distributed decision making, where the reward is determined by the maximum outcome among $K$ selected arms. When outcomes are…

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

    Bandits 中公平的代价：一个紧密的极小极大刻画

    In bandit problems, standard regret-minimizing algorithms treat exploration as an amortized cost, which can expose early participants to unfair ex-ante losses in settings such as clinical trials. Recent work addresses this by evaluating the sequence of per-round expected rewards …

  5. arXiv stat.ML TIER_1 English(EN) · Dhruv Sarkar, Soumyadeep Dutta, Sayak Ray Chowdhury ·

    Bandits 中公平的代价：一个紧密的极小极大刻画

    arXiv:2607.13402v1 Announce Type: new Abstract: In bandit problems, standard regret-minimizing algorithms treat exploration as an amortized cost, which can expose early participants to unfair ex-ante losses in settings such as clinical trials. Recent work addresses this by evalua…

  6. arXiv stat.ML TIER_1 English(EN) · Sayak Ray Chowdhury ·

    Bandits 中公平的代价：一个紧密的极小极大刻画

    In bandit problems, standard regret-minimizing algorithms treat exploration as an amortized cost, which can expose early participants to unfair ex-ante losses in settings such as clinical trials. Recent work addresses this by evaluating the sequence of per-round expected rewards …

  7. arXiv stat.ML TIER_1 English(EN) · Mo\"ise Blanchard, Dmitrii Ostrovskii, Aadirupa Saha ·

    Bandit PCA with Minimax Optimal Regret

    arXiv:2607.10936v1 Announce Type: cross Abstract: We study the bandit-feedback version of online principal component analysis (Bandit PCA): in each round $t = 1,\dots,T$, the adversary selects a $d \times d$ symmetric gain matrix $G_t$ with spectrum in $[0,1]$ and rank at most $r…

  8. arXiv stat.ML TIER_1 English(EN) · Aadirupa Saha ·

    Bandit PCA with Minimax Optimal Regret

    We study the bandit-feedback version of online principal component analysis (Bandit PCA): in each round $t = 1,\dots,T$, the adversary selects a $d \times d$ symmetric gain matrix $G_t$ with spectrum in $[0,1]$ and rank at most $r$; the learner simultaneously selects a unit vecto…

  9. arXiv stat.ML TIER_1 English(EN) · Aadirupa Saha ·

    Bandit PCA with Minimax Optimal Regret

    We study the bandit-feedback version of online principal component analysis (Bandit PCA): in each round $t = 1,\dots,T$, the adversary selects a $d \times d$ symmetric gain matrix $G_t$ with spectrum in $[0,1]$ and rank at most $r$; the learner simultaneously selects a unit vecto…

  10. Towards AI TIER_1 English(EN) · Jacob Ingle ·

    DIY 人工智能与机器学习：使用汤普森采样解决多臂老虎机问题

    <div class="medium-feed-item"><p class="medium-feed-image"><a href="https://pub.towardsai.net/diy-ai-ml-solving-the-multi-armed-bandit-problem-with-thompson-sampling-c01c27e00c68?source=rss----98111c9905da---4"><img src="https://cdn-images-1.medium.com/max/1024/0*7JFTF8U3aGqxVOSV…

相关实体

相关话题