SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for September
-
ECE Seminar: From Single-agent to Federated Reinforcement Learning
Wed, Sep 04, 2024 @ 10:00 AM - 11:00 AM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Yuejie Chi, Sense of Wonder Group Endowed Professor of Electrical and Computer Engineering in AI Systems, Carnegie Mellon University
Talk Title: From Single-agent to Federated Reinforcement Learning
Abstract: Reinforcement learning (RL) has garnered significant interest in recent years due to its success in a wide variety of modern applications. Q-learning, which seeks to learn the optimal Q-function of a Markov decision process (MDP) in a model-free fashion, lies at the heart of RL practices. However, theoretical understandings on its non-asymptotic sample complexity remain unsatisfactory, despite significant recent efforts. In this talk, we first show a tight sample complexity bound of Q-learning in the single-agent setting, together with a matching lower bound to establish its minimax sub-optimality. We then show how federated versions of Q-learning allow collaborative learning using data collected by multiple agents without central sharing, where an importance averaging scheme is introduced to unveil the blessing of heterogeneity.
Biography: Dr. Yuejie Chi is the Sense of Wonder Group Endowed Professor of Electrical and Computer Engineering in AI Systems at Carnegie Mellon University, with courtesy appointments in the Machine Learning department and CyLab. She received her Ph.D. and M.A. from Princeton University, and B. Eng. (Hon.) from Tsinghua University, all in Electrical Engineering. Her research interests lie in the theoretical and algorithmic foundations of data science, signal processing, machine learning and inverse problems, with applications in sensing, imaging, decision making, and generative AI. Among others, Dr. Chi is a recipient of the Presidential Early Career Award for Scientists and Engineers (PECASE), NSF CAREER Award, ONR YIP Award, AFOSR YIP Award, the inaugural IEEE Signal Processing Society Early Career Technical Achievement Award for contributions to high-dimensional structured signal processing, and multiple paper awards including the SIAM Activity Group on Imaging Science Best Paper Prize and IEEE Signal Processing Society Young Author Best Paper Award. She is an IEEE Fellow (Class of 2023) for contributions to statistical signal processing with low-dimensional structures.
Host: Drs. Richard M. Leahy (leahy@usc.edu) and Mahdi Soltanolkotabi (soltanol@usc.edu)
Webcast: https://usc.zoom.us/j/91569704176?pwd=zHQIlJ6vFqFmWPQYbARB8J3pXRbRiV.1Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
WebCast Link: https://usc.zoom.us/j/91569704176?pwd=zHQIlJ6vFqFmWPQYbARB8J3pXRbRiV.1
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
Semiconductors & Microelectronics Seminar - Yuanwei Li, Thursday, 9/12 at 2:30pm in EEB 132
Thu, Sep 12, 2024 @ 02:30 PM - 03:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Yuanwei Li, Stanford University
Talk Title: Towards Designing Functionality: Nano-Architected Materials for Next-Generation Sustainability and Health Monitoring
Series: Semiconductors & Microelectronics Technology
Abstract: A central aim in materials science is the ability to dictate functionality through deliberate design rules, leading to the synthesis and characterization of targeted structures. Inspired by natural materials' assembly and optical properties, my research develops nano-architected materials from nanoscale to macroscale, each tailored with specific chemical and optical properties. This talk will delve into the intersection of chemistry, nanomaterials, and optical physics to innovate materials for enhanced sustainability and health monitoring applications.My research focuses on designing functional colloidal crystals using principles inspired by the geometric intricacies observed in natural systems. Employing DNA-functionalized inorganic nanoparticles as the building blocks, I have developed multicomponent and porous colloidal crystals through programmable assemblies, advancing the complexity achievable in crystalline structures. These crystals are engineered to possess unique functionalities such as negative refraction, broadband absorption, and significant mechanical robustness. Moreover, I address synthetic challenges in creating porous crystals with tunable pore sizes ranging from 10 to 1000 nm, which can be employed in applications from advanced catalysis to optical devices like invisibility cloaks and miniaturized mechanical components.I extend my expertise to designing intricate metamaterials that synergize bottom-up assemblies with top- down lithography for health monitoring by developing optical biosensors. Focusing on the continuous, multiplexed monitoring of key metabolites associated with chronic stress, my approach integrates high-quality- factor dielectric metasurfaces with plasmonic spherical nucleic acids composed of modular DNA aptamer probes. Demonstrating sub-picomolar sensitivity, this optical sensor enables real-time, multiplexed detection across dense arrays of resonators, potentially revolutionizing portable health monitoring systems.
Biography: Yuanwei Li is a postdoctoral fellow in Materials Science and Engineering at Stanford University, as a Stanford Science Fellow under the guidance of Prof. Jennifer Dionne. She focuses on developing new optical nanomaterials for biosensing and photocatalysis. She received her PhD in Chemical and Biological Engineering at Northwestern University as a Ryan Fellow, working with Prof. Chad Mirkin. Her graduate research focused on the programmable assembly of nanoparticles into colloidal crystals with tailored chemical, optical, and mechanical properties by design. Her work has been published in Nature, Science, Nature Materials, and Science Advances. She received the MRS Graduate Student Award, Outstanding Research Award by the International Institute for Nanotechnology, the SPIE Optics and Photonics Education Scholarship, and has been named a Rising Star in Chemical Engineering by MIT.
Host: Jayakanth Ravichandran, Joshua Yang, Chongwu Zhou, Stephen Cronin, Wei Wu
More Information: Yuanwei Li Flyer.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
Audiences: Everyone Is Invited
Contact: Marilyn Poplawski
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
CSC/CommNetS-MHI Seminar: Thomas Zhang
Mon, Sep 16, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Thomas Zhang, University of Pennsylvania
Talk Title: Guarantees for Representation Learning: Distribution Shift, Optimization, Fewer Samples
Abstract: In many areas of machine learning, it is the general understanding that broadly useful features can be extracted from data across different tasks or domains. This forms the key intuition behind the "pre-train then fine-tune" paradigm, and more generally representation learning (learning the feature mappings) and transfer learning (downstream performance on unseen tasks). Naturally, there has been significant effort to document the benefit of using diverse multi-task data both empirically and theoretically. However, prior works impose various assumptions that greatly affect their applicability, especially in settings involving data generated by dynamical systems, e.g. robotics and control.
In this talk, I will introduce the multi-task representation learning problem, and walk through the pathologies arising from sequential settings, previewed in the talk title. I will then present our recent results addressing many of these issues. In particular, we provide generalization guarantees which illustrate the benefit of learning a shared representation across domains, remaining valid even when there are too few samples to solve each task individually. We then show that optimizing for the representation is surprisingly hard, requiring critical algorithmic modifications to ensure convergence. Lastly, I will show how these results, descended from iid learning, can be lifted to dynamical systems to ensure closed-loop performance.
Biography: Thomas Zhang is a 5th-year PhD student at the University of Pennsylvania advised by Prof. Nikolai Matni. His research interests involve a combination of dynamical systems, statistical learning, and control theory. Prior to Penn, Thomas received BSc’s in Mathematics and Statistics & Data Science from Yale University, where he then spent a year as a research scientist in the Applied Mathematics Program.
Host: Dr. Stephen Tu, stephen.tu@usc.edu
More Information: 2024.09.16 CSC Seminar - Thomas Zhang.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - EEB 132
Audiences: Everyone Is Invited
Contact: Miki Arlen
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
AAI-CCI-MHI Seminar on CPS
Wed, Sep 18, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Roy Fox, Assistant Professor of Computer Science Department, University of California Irvine
Talk Title: World Models -” Driven by Vision and Language; Driving Transfer
Series: EE598 Seminar Series
Abstract: An agent learning to control its environment would often do well to first model it. Compared to control policies, world models use a richer training signal and tend to generalize and transfer better. However, for a model to induce good behavior, it must be highly accurate in all reachable states, which may require too much data. Because efforts to leverage web-scale data for control are yet to succeed as they famously have for vision and language, we ask: can information encoded in vision and language foundation models help guide world modeling? In the first part of this talk, we will see two such methods: one that uses a segmentation foundation model to block visual distractions and keep state representations task-relevant; and one that queries a language model to hypothesize about abstract world models that guide exploration and planning. In the second part of the talk, we will revisit the transfer power of world models in two settings: simulation-to-reality and delayed perception. We will see how a model of a simulator can be adapted to reality with a tiny amount of data; and how a world model can transfer across varying delays of the agent's observations.
Biography: Roy Fox is an Assistant Professor of Computer Science at the University of California, Irvine. His research interests include theory and applications of control learning: reinforcement learning (RL), control theory, information theory, and robotics. His current research focuses on structured and model-based RL, language for RL and RL for language, and optimization in deep control learning of virtual and physical agents.
Host: Stephen Tu
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
Audiences: Everyone Is Invited
Contact: Ariana Perez
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
ECE Seminar: Technology development for functional and morphological imaging of the middle and inner ear
Tue, Sep 24, 2024 @ 01:30 PM - 02:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Brian E. Applegate, Professor of Otolaryngology-“Head & Neck Surgery, Ophthalmology, and Biomedical Engineering, University of Southern California
Talk Title: Technology development for functional and morphological imaging of the middle and inner ear
Abstract: Over the past decade we have been developing Optical Coherence Tomography and Vibrometry (OCTV) to measure the detailed morphology and vibratory response of the ear. With micron scale spatial resolution and subnanometer sensitivity to vibration it is well suited to measuring the spatially resolved vibratory response of both the inner and middle ear. In small animals, it is possible to image directly through the bone of the otic capsule for noninvasive spatially resolved vibrometry of the cochlear partition. In humans as well as small animals, it’s possible to image the tympanic membrane and ossicles through the ear canal to reveal the vibratory response of the middle ear. Nominally, this approach allows for the measure of vibratory response along the light path of the instrument, hence 1-D. In recent work we have developed a system that incorporates 3 separate sample arms in a single interferometer. This allows us to reconstruct the full 3-D vector of motion. We have also begun translating OCTV for use in humans with the development of hand-held and surgical microscope mounted devices which can be used in the clinic and OR. The seminar will be split between these two projects, outlining the technical design and discussing recent results for each.
Biography: Dr. Applegate is a Professor of Otolaryngology–Head & Neck Surgery, Ophthalmology, and Biomedical Engineering at the University of Southern California. He received his Ph.D. in physical chemistry from The Ohio State University. He won a National Institutes of Health postdoctoral fellowship grant to continue his training at Duke University in biomedical engineering. Upon completing his fellowship, he joined the faculty of Texas A&M University where he worked for 12 years advancing to the rank of Associate Professor of Biomedical Engineering. He moved to the University of Southern California in 2019 where he joined his longtime collaborator to continue their work on functional imaging of the ear. Throughout his career, his research has been supported by grants from the National Science Foundation, including the NSF Career award, the Department of Defense, and the National Institutes of Health. He has served on numerous study sections at the National Institutes of Health including a term on Imaging Guided Interventions and Surgery [IGIS]. He has served as an Associate Editor for IEEE Transactions on Medical Imaging and Optics Letters. He has been elected a fellow of Optica and SPIE. His research interests are broadly to develop novel biophotonic technologies and apply them to the diagnosis and monitoring of human disease.
Host: Dr. Richard M. Leahy, leahy@usc.edu
More Info: (USC NetID login required to join seminar)
Webcast: https://usc.zoom.us/j/95027937825?pwd=hEzt0iA1hkdnoINOSMiV2wrXGzcIGo.1Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
WebCast Link: https://usc.zoom.us/j/95027937825?pwd=hEzt0iA1hkdnoINOSMiV2wrXGzcIGo.1
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher
Event Link: (USC NetID login required to join seminar)
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
ECE Seminar: CMOS System-on-Chip Technology for Exploring Earth, the Solar System, and the Space
Thu, Sep 26, 2024 @ 10:00 AM - 11:00 AM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Adrian Tang, NASA Jet Propulsion Laboratory, California Institute of Technology
Talk Title: CMOS System-on-Chip Technology for Exploring Earth, the Solar System, and the Space
Abstract: In this talk, I will discuss the infusion of CMOS system-on-chip (SoC) technology into NASA-JPL instrumentation for current and future Astronomy, Earth science, and planetary science investigations. It will describe how the adoption of SoC technology has enabled a drastic reduction in the size, mass, and power consumption of space instruments, allowing them to be carried on smaller platforms while also enabling entirely new science investigations through the co-integration of mm-wave/THz and DSP elements into single-chip devices. I will discuss the fundamental design challenges these SoCs face in delivering the fidelity required for NASA’s science investigations, including sensitivity and long-term stability (Allan variance) in radiometers and high dynamic range in radar sensing. Several recently developed SoC based science instruments will be presented including: (1) The ReckTangLE mission, which carried a CMOS 183 GHz emission spectrometer and flew on a 2019 sub-orbital mission investigating stratospheric water vapor on Earth, (2) the WHATSUP 500-600 GHz spectrometer, which measures isotopic ratios of water at Europa, Titan, and Enceladus to better understand the origins of water in our solar system, (3) the Airborne Scanning Microwave Limb Sounder (ASMLS), a 340 GHz limb-sounder mission flown aboard the NASA ER-2 aircraft, (4) the SoC based ground penetrating radar (GPR) for Mars Science Helicopter that explores subsurface deposits of ice at the Martian poles to be better understand the origins of water and ice on the red planet, (5) the NASA CMOS Enhanced MetaSurface Radar mission monitoring the snowpack water content in the southwestern USA to provide the western states accurate water resource planning during periods of prolonged drought, (6) the NASA Spec-Chip instrument which explores comets and asteroids, analyzing the gasses trapped within their icy surfaces when the solar system first formed, giving us a glimpse into our cosmic origins.
Biography: Dr. Adrian Tang has over 20 years of CMOS/SiGe system-on-chip (SoC) development experience and currently directs the space system-on-chip laboratory at NASA’s Jet Propulsion Laboratory, which develops a wide range of CMOS SoCs for exploration of Earth, the solar system, and space. These SoCs are widely deployed across NASA’s spaceborne, airborne, and surface mission portfolios. He has served as a principal or co-principal investigator on over 30 NASA science and technology programs and as a principal investigator and mission manager on 3 NASA sub-orbital missions. He has authored over 160 IEEE articles in radar and mm-wave/THz remote sensing and was recently awarded the 2023 IEEE Region 6 Outstanding Engineer Award and the 2023 NASA Exceptional Achievement Medal.
Host: Dr. Mahta Moghaddam, mahta@usc.edu
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
CSC/CommNetS-MHI Seminar: Lauren Conger
Mon, Sep 30, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Lauren Conger, PhD Candidate, Control and Dynamical Systems, Caltech
Talk Title: Wasserstein Gradient Flows for Modeling Strategic Distribution Shift
Series: CSC/CommNetS-MHI Seminar Series
Abstract: We propose a novel framework for analyzing the dynamics of distribution shift in real-world systems that captures the feedback loop between learning algorithms and the distributions on which they are deployed. We propose a coupled partial differential equation model that captures fine-grained changes in the distribution over time by accounting for complex dynamics that arise due to strategic responses to algorithmic decision-making, non-local endogenous population interactions, and other exogenous sources of distribution shift. We prove convergence results in 3 settings for min-max optimization problems over measures, as well as in a cooperative setting, addressing a recent open problem posed in (Wang and Chizat 2024, Convergence of single-timescale mean-field Langevin descent-ascent for two-player zero-sum games).
Biography: Lauren Conger is a PhD candidate in Control and Dynamical Systems at Caltech, advised by Franca Hoffmann, Eric Mazumdar, and John Doyle. Her research is at the intersection of partial differential equations (PDE) analysis, game theory, and control theory. She studies systems of gradient flow PDEs through the lens of game theory with applications in machine learning, and works on system level synthesis, a new control parameterization, in a variety of contexts, including distributed optimization and PDE control.
Host: Dr. Lars Lindemann, llindema@usc.edu
Location: Hughes Aircraft Electrical Engineering Center (EEB) - EEB 132
Audiences: Everyone Is Invited
Contact: Miki Arlen
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
CSC/CommNetS-MHI Seminar: Lauren Conger
Mon, Sep 30, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Lauren Conger, PhD Candidate, Control and Dynamical Systems, Caltech
Talk Title: Wasserstein Gradient Flows for Modeling Strategic Distribution Shift
Series: CSC/CommNetS-MHI Seminar Series
Abstract: We propose a novel framework for analyzing the dynamics of distribution shift in real-world systems that captures the feedback loop between learning algorithms and the distributions on which they are deployed. We propose a coupled partial differential equation model that captures fine-grained changes in the distribution over time by accounting for complex dynamics that arise due to strategic responses to algorithmic decision-making, non-local endogenous population interactions, and other exogenous sources of distribution shift. We prove convergence results in 3 settings for min-max optimization problems over measures, as well as in a cooperative setting, addressing a recent open problem posed in (Wang and Chizat 2024, Convergence of single-timescale mean-field Langevin descent-ascent for two-player zero-sum games).
Biography: Lauren Conger is a PhD candidate in Control and Dynamical Systems at Caltech, advised by Franca Hoffmann, Eric Mazumdar, and John Doyle. Her research is at the intersection of partial differential equations (PDE) analysis, game theory, and control theory. She studies systems of gradient flow PDEs through the lens of game theory with applications in machine learning, and works on system level synthesis, a new control parameterization, in a variety of contexts, including distributed optimization and PDE control.
Host: Dr. Lars Lindemann, llindema@usc.edu
More Information: 2024.09.30 CSC Seminar - Lauren Conger.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - EEB 132
Audiences: Everyone Is Invited
Contact: Miki Arlen
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
CSC/CommNetS-MHI Seminar: Lauren Conger
Mon, Sep 30, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Lauren Conger, PhD Candidate, Control and Dynamical Systems, Caltech
Talk Title: Wasserstein Gradient Flows for Modeling Strategic Distribution Shift
Series: CSC/CommNetS-MHI Seminar Series
Abstract: We propose a novel framework for analyzing the dynamics of distribution shift in real-world systems that captures the feedback loop between learning algorithms and the distributions on which they are deployed. We propose a coupled partial differential equation model that captures fine-grained changes in the distribution over time by accounting for complex dynamics that arise due to strategic responses to algorithmic decision-making, non-local endogenous population interactions, and other exogenous sources of distribution shift. We prove convergence results in 3 settings for min-max optimization problems over measures, as well as in a cooperative setting, addressing a recent open problem posed in (Wang and Chizat 2024, Convergence of single-timescale mean-field Langevin descent-ascent for two-player zero-sum games).
Biography: Lauren Conger is a PhD candidate in Control and Dynamical Systems at Caltech, advised by Franca Hoffmann, Eric Mazumdar, and John Doyle. Her research is at the intersection of partial differential equations (PDE) analysis, game theory, and control theory. She studies systems of gradient flow PDEs through the lens of game theory with applications in machine learning, and works on system level synthesis, a new control parameterization, in a variety of contexts, including distributed optimization and PDE control.
Host: Dr. Lars Lindemann, llindema@usc.edu
More Information: 2024.09.30 CSC Seminar - Lauren Conger.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - EEB 132
Audiences: Everyone Is Invited
Contact: Miki Arlen
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor.
