M
Mingda Qiao
Researcher at Stanford University
Publications - 25
Citations - 292
Mingda Qiao is an academic researcher from Stanford University. The author has contributed to research in topics: Upper and lower bounds & Computer science. The author has an hindex of 9, co-authored 22 publications receiving 211 citations. Previous affiliations of Mingda Qiao include Tsinghua University.
Papers
More filters
Posted Content
On Generalization Error Bounds of Noisy Gradient Methods for Non-Convex Learning
Jian Li,Xuanyuan Luo,Mingda Qiao +2 more
TL;DR: A new framework, termed Bayes-Stability, is developed for proving algorithm-dependent generalization error bounds for learning general non-convex objectives and it is demonstrated that the data-dependent bounds can distinguish randomly labelled data from normal data.
Proceedings Article
Nearly Instance Optimal Sample Complexity Bounds for Top-k Arm Selection.
Lijie Chen,Jian Li,Mingda Qiao +2 more
TL;DR: A novel complexity term is obtained to measure the sample complexity that every Best-$k$-Arm instance requires and an elimination-based algorithm is provided that matches the instance-wise lower bound within doubly-logarithmic factors.
Proceedings Article
Collaborative PAC Learning
TL;DR: A collaborative PAC learning model, in which k players attempt to learn the same underlying concept, with an Omega(ln(k)) overhead lower bound, showing that the results are tight up to a logarithmic factor.
Proceedings Article
Towards Instance Optimal Bounds for Best Arm Identification
Lijie Chen,Jian Li,Mingda Qiao +2 more
TL;DR: For instance, Chen and Li as discussed by the authors showed that for any Gaussian best-arm instance with gaps of the form 2-k, there is a monotone algorithm with sample complexity O(n, √ n, polylog n, log n, ε 2 n, ∵ 2 log n) for best-armed instances.
Proceedings Article
Nearly Optimal Sampling Algorithms for Combinatorial Pure Exploration
TL;DR: In this article, the authors studied the combinatorial pure exploration problem of best-set in stochastic multi-armed bandits, where the goal is to identify the feasible subset with the maximum total mean using as few samples as possible.