Tutorial at IFIP Performance 2023 on Stochastic Matching Networks: Theory and Applications - two part tutorial jointly with Sushil Varma. Slides of both the parts are here.
Part 1 presents a heavy traffic theory of matching queues, and illustrates their use of pricing and matching in online platforms consisting of a bipartite graph of matching queues. Part 2 by Sushil presents multiway matchings and their use in Quantum switches, and also a thorough literature survey of matching queues.
Tutorial at INFORMS 2023 in APS Cluster on Stochastic Approximation of Contractive Operators: Finite-Time Guarantees and Applications in Reinforcement Learning. Slides are here and a draft of a survey paper is here.
A Tutorial on Finite-Sample Guarantees of Contractive Stochastic Approximation With Applications in Reinforcement Learning - two part tutorial at Simons Institute in Oct 2022 conducted jointly with Zaiwei Chen. Videos and slides of Part 1 and Part 2.
Part 1 gives an overview of Stochastic Approximation of general contractive operators. Part 2 uses these results to obtain finite-sample bounds of various RL algorithms.
A three hour 'Tutorial on Stein's method' delivered as a guest lecture in Prof. Yash Kanoria's course at Columbia Univeristy. Lecture Notes and Video.
This tutorial is in three parts. Part 1 is an introduction and overview of Stein's Method for Normal Approximation. Part 2 illustrates an application of the method by proving the rate of convergence in the central limit theorem, i.e., the Berry-Esseen Theorem. Part 3 illustrates an application on exponential approximation by studying a single server queue in Heavy-Traffic.
Virtual Tutorial at Sigmetrics 2021 on 'Lyapunov Drift Methods for Stochastic Recursions: Optimization, Reinforcement Learning and Resource Allocation,' conducted jointly with Zaiwei Chen. Slides and Videos: Part 1, Part 2, Part 3.
Part 1 provides on overview of drift methods for Stochastic Recursions with two illustrative examples - one in queues and the second in stochastic Approximation. Part 2 by Zaiwei, generalizes the drift arguments to study general contractive stochastic approximation with applications in Reinforcement Learning. Part 3 presents drift arguments in Stochastic Processing Networks, with a focus on data center networks and input-queued switch.
Tutorial at EC 2019 on 'Backpressure: A Network Control Approach to Running Marketplaces.' In this four part tutorial, I conducted Part 2 on 'Drift Method and Transform Techniques for Performance Analysis.' Video, Slides.
Tutorial at IFIP Performance 2017 on 'The Drift Method for Heavy Traffic Limits, with Applications in Data Centers and Networks,' conducted jointly with R. Srikant and Weina Wang in three parts. Slides.
Virtual Keynote at Young European Queueing Theorists Workshop in June 2021, on 'A Heavy-Traffic Theory of Two-Sided Queues with Applications in Matching Platforms and Online Marketplaces' - Video, Slides.
Virtual Seminar in Communications and Signal Processing Seminar Series at University of Michigan, on 'A Lyapunov Theory for Finite-Sample Guarantees of Asynchronous, Value-Based Reinforcement Learning Algorithms,' February 2021. Slides and Video.
Virtual Seminar in UT Austin's ML seminar series on 'Finite Sample Convergence Bounds of Off-Policy RL Algorithms,' October 2020. Slides and Video.
Invited talk on 'Delay-Optimal Scheduling for Data Center Networks and Input-Queued Switches in Heavy Traffic,' at the twelfth International Conference on Stochastic Networks in June 2016. Slides.
Here is a list of selected seminar series, primarily at Georgia Tech, along with info about how to join the mailing lists. This list is designed so that new students and postdocs can quickly subscribe to these mailing lists in order to be aware of all the seminars going on.
ISyE Seminar Series. Join the mailing list here.
The same list also announces ISyE student seminars.
SNAPP virtual seminar series. Join this Google group or send an email to snappseminar+subscribe@googlegroups.com
Machine Learning Seminar Series. Join mailing list here.
ARC Colloquium. Join mailing list here.
ACO Student Seminar. Join mailing list here.
Math department has several seminar series including the Stochastics Seminar Series. See this page for info about joining the mailing lists.
Seminars not active anymore (but previous talks may be accessible for some)
Theoretical Neuroscience Seminar - A special series in Spring 2023 sponsored by IDEaS
IDEaS AI seminar series. Join here.
DCL Seminar series. Join mailing list here or by sending an email to sympa@lists.gatech.edu with subject 'subscribe dcl-students'