题目: Algorithms with Distributional Predictions

主讲人：Michael Dinitz 副教授

时间：2026年09月12日（周六）9:00-10:00

地点：腾讯会议 830-741-764

主办单位：理学院

主讲人简介:

Michael Dinitz是约翰斯・霍普金斯大学（Johns Hopkins University）计算机科学系副教授，同时在应用数学与统计系兼任教职，隶属于该校数据科学数学研究所、数据科学与人工智能研究所。他于卡内基梅隆大学取得博士学位，曾在魏茨曼科学研究所从事博士后研究。他的研究方向以算法为主，重点研究近似算法、图算法、分布式算法以及计算机网络理论。主持多项美国国家科学基金会（NSF）项目。发表期刊论文及会议论文等80余篇。

摘要：

Algorithms with (machine-learned) predictions is a powerful framework for combining traditional worst-case algorithms with modern machine learning. However, the vast majority of work in this space assumes that the prediction itself is non-probabilistic, even if it is generated by some stochastic process (such as a machine learning system). This is a poor fit for modern ML, particularly modern neural networks, which naturally generate a distribution.  We extend the predictions framework to handle distributional predictions, and show that this richer setting allows for strong algorithms and bounds even for classical problems such as search, ski rental, and contention resolution.