# Constant-Stepsize Stochastic Approximation: Bias and Steady-State Convergence (Qiaomin Xie, University of Wisconsin–Madison) - **When:** Mon, Nov 23, 2026, 3:30 PM PST – 4:30 PM PST - **Where:** 3108 Etcheverry Hall, Berkeley - **Price:** Free - **Organizer:** [Department of Industrial Engineering and Operations Research](https://eventing.ing/organizers/orgz_1m4v68qxrf9kcmyted11fy) - **View on UC Berkeley:** https://events.berkeley.edu/IEOR/event/337386-constant-stepsize-stochastic-approximation-bias-and-s Title: Constant-Stepsize Stochastic Approximation: Bias and Steady-State Convergence Abstract: Stochastic approximation (SA) underpins many algorithms in machine learning and reinforcement learning, such as stochastic gradient descent (SGD), TD learning, and Q-learning. Constant-stepsize SA is widely used in practice for its simplicity and fast initial progress. Viewing the joint data–iterate sequence as a Markov chain, we establish its geometric convergence to a… Source: UC Berkeley. Eventing indexes this event and links to the source for tickets and the full description. Same event as JSON: https://api.eventing.ing/events/evnt_1m6rdkkq01r9rtsx9tf29m