Efficient Nonmyopic Active Search
Efficient Nonmyopic Active Search
复制标题
高效的非近视主动搜索
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
R. Garnett
中科院分区：
文献类型：
--
作者：
Shali Jiang;Gustavo Malkomes;Geoffrey A. Converse;Alyssa Shofner;Benjamin Moseley;R. Garnett
主动搜索是一种学习范式，其目标是尽可能多地主动识别给定类别的成员。许多现实世界中的问题都可以归结为主动搜索，包括药物研发、欺诈检测以及产品推荐。先前的研究已经推导出了针对该问题的贝叶斯最优策略，但遗憾的是，由于其指数级的复杂度，该策略难以实际应用。在实际操作中，人们转而使用近视近似法，即在决策过程中仅向前看少数几步（例如1 - 3步）。我们提出了一种新颖的主动搜索策略，该策略始终考虑整个剩余预算，因此并非近视策略，同时仍保持高效。我们的方法能够自动且动态地以与预算相符的方式平衡探索与利用，而无需依赖权衡参数。我们还开发了一种边界界定技术，以便在使用某些自然概率模型时实现更高的效率。实验结果表明，相较于最优策略的近视近似法，我们的方法表现更优。
Active search is a learning paradigm with the goal of actively identifying as many members of a given class as possible. Many real-world problems can be cast as an active search, including drug discovery, fraud detection, and product recommendation. Previous work has derived the Bayesian optimal policy for the problem, which is unfortunately intractable due to exponential complexity. In practice, myopic approximations are used instead, only looking a small number (e.g., 1–3) of steps ahead in the decision process. We propose a novel active search policy that always considers the entire remaining budget and is thus nonmyopic, yet remains efficient. Our approach automatically and dynamically balances exploration and exploitation in a manner consistent with the budget, without relying on a tradeoff parameter. We also develop a bounding technique to achieve greater efficiency when using certain natural probability models. Experimental results show superior performance of our method over myopic approximations to the optimal policy.
引入迭代虚拟筛选的“主动搜索”方法
影响因子：
3.5
作者：
Roman Garnett;Thomas Gärtner;Martin Vogt;Jürgen Bajorath
通讯作者：
Jürgen Bajorath
高效非近视批量主动搜索
DOI：
--
发表时间：
2018
期刊：
32nd Conference on Neural Information Processing Systems (NeurIPS 2018
影响因子：
--
作者：
Jiang, Shali;Malkomes, Gustavo;Abbott, Matthew;Moseley, Benjamin;Garnett, Roman
通讯作者：
Garnett, Roman