Seminars

Friday, April 1, 2022
4:10 – 5:00 p.m.
Zoom: https://tamu.zoom.us/j/96343481647
 
Dr. Siva Theja Maguluri
Assistant Professor in the H. Milton Stewart School of Industrial and Systems Engineering, Georgia Tech

Talking Points

• Lyapunov methods for Stochastic Approximation and Reinforcement learning
• Generalized Moreau Envelop based on infimal convolution smoothing as a Lyapunov function for Stochastic Approximation of contractive operators
• Unified framework to obtain sample complexity of a large class of RL algorithms
• Linear speedup in the number of agents for Federated Reinforcement learning

Title
“A Lyapunov Theory of Finite-Sample Guarantees of Stochastic Approximation and Reinforcement Learning”

Abstract
The focus of our work is to obtain finite-sample and/or finite-time convergence bounds of various model-free Reinforcement Learning (RL) algorithms. Many RL algorithms involve solving the Bellman fixed point equation, which is done using Stochastic Approximation (SA). SA is a popular approach for solving fixed point equations when the information is corrupted by noise. We develop a Lyapunov framework and obtain mean square error bounds on the convergence of a general class of SA algorithms for contractive operators under general norms and Markovian noise. The key tool we use is generalized Moreau envelope as a smooth potential/ Lyapunov function. These powerful results immediately provide sample complexity results of a large class of RL algorithms including TD learning, Q-learning, actor-critic algorithms, their off-policy variants, and their distributed variants. The talk will present a couple of these applications in off-policy RL and/or Federated RL.

Biography
Siva Theja Maguluri is Fouts Family Early Career Professor and Assistant Professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech. He obtained his Ph.D. and MS in ECE as well as MS in Applied Math from UIUC, and B.Tech in Electrical Engineering from IIT Madras. His research interests span the areas of Control, Optimization, Algorithms and Applied Probability. In particular, he works on Reinforcement Learning theory, scheduling, resource allocation and revenue optimization problems that arise in a variety of systems including Data Centers, Cloud Computing, Wireless Networks, Block Chains, Ride hailing systems, etc. His research and teaching are recognized through several awards including the  “Best Publication in Applied Probability” award, NSF CAREER award, second place award at INFORMS JFIG best paper competition, Student best paper award at IFIP Performance, “CTL/BP Junior Faculty Teaching Excellence Award,” and “Student Recognition of Excellence in Teaching: Class of 1934 CIOS Award.”

Dr. Maguluri: https://sites.google.com/site/sivatheja/

On Zoom @ 4:10 p.m. on Friday, 4/1/22

Join Zoom Meeting
https://tamu.zoom.us/j/96343481647
Meeting ID: 963 4348 1647

Friday, March 4, 2022
4:10 – 5:00 p.m.
Zoom: https://tamu.zoom.us/j/96343481647
 
Dr. Tommaso Melodia
William L. Smith Professor, Northeastern University

Talking Points
– Open, Programmable, and Virtualized Wireless Architectures
– O-RAN architecture: Research Challenges
– Deep Reinforcement Learning for sequential decision making in the Open RAN
– Resources available for at scale prototyping, testing, data collection: PAWR, Colosseum

Title
“Toward AI-based Control and Orchestration in the Open RAN: Architectures, Algorithms, Testbeds”

Abstract
This talk will present an overview of our work on laying the basic principles to design open, programmable, AI-driven, and virtualized next-generation wireless networks. We will cover in detail challenges and opportunities associated with the evolution of cellular systems into cloud-native softwarized architectures enabling fine grained AI-based control of end-to-end functionalities on mobile devices, in the radio access network, and at the edge. We will also discuss existing testbeds and platforms available to the community for prototyping, experimentation, and data collection in virtualized and softwarized wireless systems.

Biography
Tommaso Melodia is the William Lincoln Smith Professor with the Department of Electrical and Computer Engineering at Northeastern University in Boston. He is also the Founding Director of the Institute for the Wireless Internet of Things and the Director of Research for the PAWR Project Office. He received his Laurea (integrated BS and MS) from the University of Rome – La Sapienza and his Ph.D. in Electrical and Computer Engineering from the Georgia Institute of Technology in 2007. He is an IEEE Fellow and recipient of the National Science Foundation CAREER award. Prof. Melodia is serving as Editor in Chief for Computer Networks, and has served as Associate Editor for IEEE Transactions on Wireless Communications, IEEE Transactions on Mobile Computing, IEEE Transactions on Multimedia, among others. He was the Technical Program Committee Chair for IEEE Infocom 2018, and General Chair for ACM MobiHoc 2020, IEEE SECON 2019, ACM Nanocom 2019, and ACM WUWNet 2014. Prof. Melodia’s research on modeling, optimization, and experimental evaluation of wireless networked systems has been funded by US federal agencies and industry.

Dr. Melodia: https://ece.northeastern.edu/wineslab/tmelodia.php

On Zoom @ 4:10 p.m. on Friday, 3/4/22

Join Zoom Meeting
https://tamu.zoom.us/j/96343481647
Meeting ID: 963 4348 1647

Friday, February 25, 2022
4:10 – 5:00 p.m.
Zoom: https://tamu.zoom.us/j/96343481647
 
Dr. Adam Wierman
Professor of Computing and Mathematical Sciences and Director of Information Sciences and Technology, Caltech

Title
“Online Optimization and Control using Black-Box Predictions”

Abstract
Making use of modern black-box AI tools is potentially transformational for online optimization and control. However, such machine-learned algorithms typically do not have formal guarantees on their worst-case performance, stability, or safety. So, while their performance may improve upon traditional approaches in “typical” cases, they may perform arbitrarily worse in scenarios where the training examples are not representative due to, e.g., distribution shift or unrepresentative training data. This represents a significant drawback when considering the use of AI tools for energy systems and autonomous cities, which are safety-critical. A challenging open question is thus: Is it possible to provide guarantees that allow black-box AI tools to be used in safety-critical applications? In this talk, I will introduce recent work that aims to develop algorithms that make use of black-box AI tools to provide good performance in the typical case while integrating the “untrusted advice” from these algorithms into traditional algorithms to ensure formal worst-case guarantees. Specifically, we will discuss the use of black-box untrusted advice in the context of online convex body chasing, online non-convex optimization, and linear quadratic control, identifying both novel algorithms and fundamental limits in each case.

Biography
Adam Wierman is a Professor in the Department of Computing and Mathematical Sciences at Caltech. He received his Ph.D., M.Sc., and B.Sc. in Computer Science from Carnegie Mellon University and has been a faculty at Caltech since 2007. Adam’s research strives to make the networked systems that govern our world sustainable and resilient. He is best known for his work spearheading the design of algorithms for sustainable data centers and his co-authored book on “The Fundamentals of Heavy-tails”. He is a recipient of multiple awards, including the ACM Sigmetrics Rising Star award, the ACM Sigmetrics Test of Time award, the IEEE Communications Society William R. Bennett Prize, multiple teaching awards, and is a co-author of papers that have received “best paper” awards at a wide variety of conferences across computer science, power engineering, and operations research.

For more on Dr. Wierman see https://adamwierman.com.

On Zoom @ 4:10 p.m. on Friday, 2/25/22

Join Zoom Meeting
https://tamu.zoom.us/j/96343481647
Meeting ID: 963 4348 1647

Friday, February 11, 2022
4:10 – 5:00 p.m.
Zoom: https://tamu.zoom.us/j/96343481647
 
Dr. Dinesh Bharadia
Asst. Professor, Center for Wireless Communications, University of California San Diego

Title
“Wireless sensing meets autonomous systems: Deployable Low-Overhead RF Localization and Mapping”

Abstract
The ability to track and locate devices and robots has become increasingly important for new and existing applications such as tracking Virtual Reality headsets, Robotics, or just your smartphone. For such purposes, in the last two decades, RF localization has gained traction, however, without any real-world deployment. Primarily due to lack of low-overhead deployment and being accurate and robust. I will present new designs for localization systems that leverage novel hardware and software capabilities to push us closer to this vision.

Specifically, I will present an autonomous robot system as a platform to enable us closer to this vision:

1. I will discuss enabling context-assisted accurate indoor localization, which leverages the autonomous robot to collect large-scale data sets and mapping. We develop techniques to enable accurate indoor localization within the context of a given map overcoming the shortcomings of the corner cases and providing reliable and robust location estimates.
2. I will discuss the mapping requirements for these indoor environments that provide context for the predicted location. I will present a system that accurately maps the WiFi anchors in the given physical map to a few millimeters.
3. Finally, I will briefly present how robotics can use WiFi as a sensor like LiDar for SLAM algorithms.

Furthermore, my group has enabled low-power and low-latency localization systems for indoor scenarios ranging from home, office to warehouse to meet next-generation Industrial IoT applications, localizing Bluetooth and UWB devices. All of this work can be found under wcsng.ucsd.edu/localization.html

Biography
Dinesh Bharadia has been faculty in ECE at the University of California San Diego since January 2018. He received his Ph.D. from Stanford University in 2016 and was a Postdoctoral Associate at MIT. Specifically, in his dissertation, he built a prototype of a radio that invalidated a long-held assumption in wireless is that radios cannot transmit and receive at the same time on the same frequency. From 2013 to 2015, he was a Principal Scientist for Kumu Networks. He commercialized his research on full-duplex radios, building a product that underwent successful field trials at Tier 1 network providers worldwide like Deutsche Telekom and SK Telecom. This product is currently under deployment. He is the principal advisor for multiple startups developed atop his research. His research interests include advancing the theory and design of modern wireless communication systems, wireless imaging, sensor networks, and wireless sensing, and broadly sensing and networking. Recently, he has been actively involved in designing cooperative autonomous systems, specifically in autonomous perception.

In recognition of his work, Dinesh was named to Forbes 30 under 30 for the science category worldwide list. Dinesh was also named a Marconi Young Scholar for outstanding wireless research and awarded the Michael Dukakis Leadership award. He was also named as one of the top 35 Innovators under 35 in the world by MIT Technology Review in 2016. Dinesh is also the recipient of the Sarah and Thomas Kailath Stanford Graduate Fellowship.

Dr. Bharadia Google Scholar Page: https://scholar.google.com/citations?user=5SjaXJsAAAAJ

On Zoom @ 4:10 p.m. on Friday, 2/18/22

Join Zoom Meeting
https://tamu.zoom.us/j/96343481647
Meeting ID: 963 4348 1647

Friday, February 18, 2022
4:10 – 5:00 p.m.
Zoom Recording
 
Dr. Gauri Joshi
Asst. Professor, Dept. of Electrical & Computer Engineering, Carnegie Mellon University

Title
“Tackling Computational Heterogeneity in Federated Learning”

Talking Points

• Data Heterogeneity in Federated Learning
• Variable Computation Speeds of Clients Participating in Federated Learning
• Client Selection in Federated Learning

Abstract
The future of machine learning lies in moving both data collection as well as model training to the edge. The emerging area of federated learning seeks to achieve this goal by orchestrating distributed model training using a large number of resource-constrained mobile devices that collect data from their environment. Due to limited communication capabilities as well as privacy concerns, the data collected by these devices cannot be sent to the cloud for centralized processing. Instead, the nodes perform local training updates and only send the resulting model to the cloud. A key aspect that sets federated learning apart from data-center-based distributed training is the inherent heterogeneity in data and local computation at the edge clients. In this talk, I will present our recent work on tackling computational heterogeneity in federated optimization, firstly in terms of heterogeneous local updates made by the edge clients, and secondly in terms of intermittent client availability.

Biography
Gauri Joshi is an assistant professor in the ECE department at Carnegie Mellon University since September 2017. Previously, she worked as a Research Staff Member at IBM T. J. Watson Research Center. Gauri completed her Ph.D. from MIT EECS in June 2016, advised by Prof. Gregory Wornell. She received her B.Tech and M.Tech in Electrical Engineering from the Indian Institute of Technology (IIT) Bombay in 2010. Her awards and honors include the NSF CAREER Award (2021), ACM Sigmetrics Best Paper Award (2020), NSF CRII Award (2018), IBM Faculty Research Award (2017), Best Thesis Prize in Computer science at MIT (2012), and Institute Gold Medal of IIT Bombay (2010).

Dr. Joshi’s Website: https://www.andrew.cmu.edu/user/gaurij/

On Zoom @ 4:10 p.m. on Friday, 2/18/22

Join Zoom Meeting
https://tamu.zoom.us/j/96343481647
Meeting ID: 963 4348 1647