212 citations · 419 across the 46 of their papers we have counts for
7 papers · 1 filter
A Primal-Dual Approach to Solving Variational Inequalities with General Constraints
Tatjana Chavdarova, Tong Yang, Matteo Pagliardini +1
Yang et al. (2023) recently showed how to use first-order gradient methods to solve general variational inequalities (VIs) under a limiting assumption that analytic solutions of sp…
Continuous-time Analysis for Variational Inequalities: An Overview and Desiderata
Tatjana Chavdarova, Ya-Ping Hsieh, Michael I. Jordan
Algorithms that solve zero-sum games, multi-objective agent objectives, or, more generally, variational inequality (VI) problems are notoriously unstable on general problems. Owing…
Solving Constrained Variational Inequalities via a First-order Interior Point-based Method
Tong Yang, Michael I. Jordan, Tatjana Chavdarova
We develop an interior-point approach to solve constrained variational inequality (cVI) problems. Inspired by the efficacy of the alternating direction method of multipliers (ADMM)…
Geometric Methods for Sampling, Optimisation, Inference and Adaptive Agents
Alessandro Barp, Lancelot Da Costa, Guilherme França +4
In this chapter, we identify fundamental geometric structures that underlie the problems of sampling, optimisation, inference and adaptive decision-making. Based on this identifica…
A Statistical Analysis of Polyak-Ruppert Averaged Q-learning
Xiang Li, Wenhao Yang, Jiadong Liang +2
We study Q-learning with Polyak-Ruppert averaging in a discounted Markov decision process in synchronous and tabular settings. Under a Lipschitz condition, we establish a functiona…
Parallelizing Contextual Bandits
Jeffrey Chan, Aldo Pacchiano, Nilesh Tripuraneni +3
Standard approaches to decision-making under uncertainty focus on sequential exploration of the space of decisions. However, \textit{simultaneously} proposing a batch of decisions,…
First-order Methods Almost Always Avoid Saddle Points
Jason D. Lee, Ioannis Panageas, Georgios Piliouras +3
We establish that first-order methods avoid saddle points for almost all initializations. Our results apply to a wide variety of first-order methods, including gradient descent, bl…