Select a calendar:
Filter October Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for October
-
ECE Seminar: Unlocking the Future: Designing Next-Generation AI Chips with AI Algorithms
Tue, Oct 01, 2024 @ 10:00 AM - 11:00 AM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Sung Kyu Lim, Motorola Solutions Foundation Professor, Georgia Institute of Technology
Talk Title: Unlocking the Future: Designing Next-Generation AI Chips with AI Algorithms
Abstract: Data, chips, and algorithms form the backbone of the AI revolution, demanding hardware as sophisticated as orchestrating a bustling city. In this technological realm, GPUs and high-bandwidth memory are essential yet frequently strained by the immense volume of data traffic. By employing 2.5D and 3D IC architectures through heterogeneous integration, we can greatly enhance energy efficiency and reduce latency in data transfers. A key component of this advancement is the automation of design and simulation for heterogeneous AI chips, where powerful algorithms take the lead, rather than humans. This remarkable capability hinges on advanced electronic design automation (EDA) tools. At Georgia Tech, my research team merges AI-driven and traditional algorithms to bolster EDA capabilities, specifically engineered for developing cutting-edge heterogeneous AI chips. In my talk, I will spotlight these innovations and address the ongoing challenges in AI chip design and EDA.
Biography: Prof. Sung Kyu Lim earned his Ph.D. in Computer Science from UCLA in 2000. Since 2001, he has been a faculty member at the School of Electrical and Computer Engineering at the Georgia Institute of Technology. His research explores the architecture, design, and electronic design automation (EDA) of 2.5D and 3D integrated circuits, contributing to over 400 published papers. He received the Best Paper Awards from the IEEE Transactions on CAD in 2022 and the ACM Design Automation Conference in 2023. He is an IEEE Fellow and served as a program manager at DARPA's Microsystems Technology Office from 2022 to 2024.
Host: Dr. Peter Beerel, pabeerel@usc.edu
More Info: (USC NetID login required to join seminar)
Webcast: https://usc.zoom.us/j/98539005883?pwd=naX0FZKrFLJwk7umPV6nneLbvRzZQF.1Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132 - LOCATION CHANGE
WebCast Link: https://usc.zoom.us/j/98539005883?pwd=naX0FZKrFLJwk7umPV6nneLbvRzZQF.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. -
MHI - Physics Joint Seminar Series, Haocun Yu, Tuesday, October 1st at 2pm in EEB 248 & Zoom
Tue, Oct 01, 2024 @ 02:00 PM - 03:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Haocun Yu, Marie-Curie Postdoctoral Fellow - University of Vienna
Talk Title: Using Quantum Optics to Illuminate the Universe's Mysteries
Series: MHI Physics Joint Seminar Series
Abstract: Advanced quantum techniques are revolutionizing our ability to observe and understand the universe. From employing squeezing in LIGO detectors to demonstrate human-scale macroscopic quantum phenomena, to utilizing photon-counting methods for measuring Earth's rotation and detecting dark matter, I will discuss how quantum optical applications enhance precision measurements, interface quantum mechanics and gravity, and offer new insights into fundamental questions about the nature of our universe.
Biography: Haocun Yu is a Marie-Cuire Postdoctoral Fellow at the University of Vienna working with Prof. Philip Walther. She completed her Ph.D. in physics in MIT LIGO group working with Prof. Nergis Mavalvala, working on quantum techniques and phenomena for gravitational-wave detectors. Her research interests lie in using various quantum techniques and precision sensing methods for fundamental physics. Her work has been recognized with honors including the MIT Martin Deutsch Award, APS Carl E. Anderson Dissertation Award, and Boeing Quantum Creators Prize. She is enthusiastic about continuing interdisciplinary work that advances quantum technologies and addresses intriguing fundamental questions about our world.
Host: Quntao Zhuang, Eli Levinson-Falk, Jonathan Habif, Daniel Lidar, Kelly Luo, Todd Brun, Tony Levi, Stephan Haas
Webcast: https://usc.zoom.us/j/92584409725More Information: Haocun Yu New Flyer.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132 - Location Change
WebCast Link: https://usc.zoom.us/j/92584409725
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: Laurent Lessard
Mon, Oct 07, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Laurent Lessard, Associate Professor, Mechanical and Industrial Engineering } Northeastern University
Talk Title: An automatic system to detect equivalence between iterative algorithms
Series: CSC/CommNetS-MHI Seminar Series
Abstract: Large-scale optimization problems in machine learning, signal processing, multi-agent systems, and imaging have fueled ongoing interest in iterative optimization algorithms. New optimization algorithms are regularly proposed in order to capture more complicated models, reduce computational burdens, or obtain stronger performance and convergence guarantees. But how can we be sure a recently proposed algorithm is novel? Algorithms can be written in different equivalent ways that are not always obvious, and with optimization being increasingly prevalent across different applications, popular algorithms are routinely "re-discovered". In this talk, we present a framework for reasoning about equivalence of iterative algorithms. Our framework is based on concepts from control theory and linear systems theory and can identify equivalence for a variety of algorithm classes: (a) single-oracle algorithms such as gradient-based methods, (b) multi-oracle algorithms such as distributed optimization algorithms, primal-dual methods, and operator-splitting methods, and (c) algorithms that use different but related oracles, such as subdifferentials, proximal operators, and Fenchel conjugates. Our work is a promising step towards an integrated and principled methodology for analyzing and designing control systems that use optimization algorithms "in the loop".
Biography: Laurent Lessard is an Associate Professor of Mechanical and Industrial Engineering at Northeastern University, Boston, USA, and a core faculty member of the Experiential Institute for AI. He received a B.A.Sc. in Engineering Science from the University of Toronto, and the M.S. and Ph.D. in Aeronautics and Astronautics at Stanford University. His research interests include: decentralized control, robust control, optimization, and machine learning. Before joining Northeastern, he was a Charles Ringrose Assistant Professor of Electrical and Computer Engineering at the University of Wisconsin–Madison. Prior to that, he was an LCCC Postdoc in the Department of Automatic Control at Lund University, Sweden, and a postdoctoral researcher in the Berkeley Center for Control and Identification at the University of California, Berkeley. Laurent is a recipient of the Hugo Schuck best paper award and the NSF CAREER award. He is also a Senior Member of IEEE.
Host: Dr. Lars Lindemann, llindema@usc.edu
More Information: 2024.10.07 CSC Seminar - Laurent Lessard.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, Oct 09, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Shai Revzen, Associate Professor Department of Electrical & Computer Science University of Michigan
Talk Title: More legs are different: the surprising simplicity of multi-legged locomotion
Series: EE598 Seminar Series
Abstract: Most of the animals that move with legs in the world do so with six or more legs, yet humans have focused primarily on bipeds and quadrupeds in designing legged robots. This talk will present some theoretical and experimental results that suggest that multi-legged robots with six or more legs exhibit some surprising properties that challenge our anthropocentric intuitions about locomotion. Modeling multi-legged motion fairly accurately, at single percentage points of relative error, turns out to be much easier than naively expected. This is both due to event-selected hybrid systems resolving multi-contact collisions in a smooth way, and due to the surprisingly high accuracy of geometric mechanics models on dry friction problems to which they shouldn't really apply. Together our results suggest that modeling and learning how to move with many legs might be much easier than has previously been thought.
Biography: Shai Revzen is an Associate Professor of Electrical Engineering and Computer Science in the University of Michigan's College of Engineering, and holds a courtesy faculty appointment in the Department of Ecology and Evolutionary Biology. He received his PhD in Integrative Biology doing research in the PolyPEDAL Lab at the University of California at Berkeley, and did his postdoctoral work in the GRASP Laboratory of the University of Pennsylvania. Prior to his academic work, Shai spent a decade in the tech industry, rising to Chief Architect R&D of the convergent systems division of Harmonic Lightwaves (HLIT). He is currently co-founder and Chief Science Officer of Acculine Medical, and General Manager of his consulting company, Izun, Inc. In his spare time he does martial arts and studies for a JD Law degree at Wayne State University.
Host: Feifei Qian
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. -
CSC/CommNetS-MHI Seminar: Verena Häberle
Mon, Oct 14, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Verena HaÌberle, PhD student, Automatic Control Laboratory | ETH Zurich, Switzerland
Talk Title: Virtual Power Plant Control for Dynamic Ancillary Services Provision
Series: CSC/CommNetS-MHI Seminar Series
Abstract: This presentation focuses on innovative control strategies for dynamic virtual power plants (DVPPs) aimed at providing dynamic ancillary services efficiently. The first part highlights the importance of heterogeneity among distributed energy resources in reliably delivering services like fast frequency and voltage control across various power and energy levels. A "divide-and-conquer" approach, along with dynamic participation factors and local matching controllers, is proposed. The second part introduces a closed-loop strategy incorporating data-driven techniques to adapt ancillary services to local grid conditions. Structural encoding of dynamic ancillary services and a "perceive-and-optimize" strategy ensure stable and optimal performance while meeting grid-code and device-level requirements. Numerical case studies and hardware experiments validate the effectiveness of these approaches, promising improved grid stability and efficiency.
Biography: Verena Häberle is a Ph.D. student at the Automatic Control Laboratory, ETH Zurich, Switzerland, working under the supervision of Prof. Florian Dörfler since June 2020. She earned both her B.Sc. and M.Sc. degrees in Electrical Engineering and Information Technology from ETH Zurich in 2018 and 2020, respectively. Since Sept 2024, she is a visiting student researcher with the Netlab group at the California Institute of Technology (CALTECH), supervised by Prof. Steven Low. Her research focuses on dynamic ancillary services provision, control design for dynamic virtual power plants, and data-driven converter control for future power systems.
Host: Dr. Lars Lindemann, llindema@usc.edu
More Information: 2024.10.14 CSC Seminar - Verena Häberle.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. -
A.V. Balakrishnan Awards Ceremony - Dr. Earl H. Dowell
Wed, Oct 16, 2024 @ 02:00 PM - 04:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Earl H. Dowell, William Holland Hall Professor of the Pratt School of Engineering, Duke University
Talk Title: Fluid Structural Thermal Interaction (FSTI) in Hypersonic Flow
Abstract: When flowing fluids and deformable structures interact, they may become unstable (flutter) and if the system is nonlinear this may lead to limit cycle oscillations and even chaotic dynamics. Physical phenomena of interest include wind induced oscillations of long span bridges and tall buildings, internal flows in nuclear reactors and gas turbines, blood flow through arteries and airflow over human tongues. However historically and even today much of the progress is driven by aerospace applications including high performance flight vehicles be they aircraft, jet engines, launch vehicles, missiles or rotorcraft. Current interest in FSTI in hypersonic flow is high and will be the subject of this talk. Both experimental and theoretical (computational) work will be discussed.
Event Program
Reception 2:00PM - 2:30PM
Remarks 2:35PM - 3:15PM
Awardee Lecture 3:15PM - 4:00PM
Award Presentation 4:00PM - 4:15PM
Biography: Dr. Dowell is an elected member of the National Academy of Engineering, an Honorary Fellow of the American Institute of Aeronautics and Astronautics (AIAA) and a Fellow of the American Academy of Mechanics and the American Society of Mechanical Engineers. He has also served as Vice President for Publications and member of the Executive Committee of the Board of Directors of the AIAA; as a member of the United States Air Force Scientific Advisory Board; the Air Force Studies Board, the Aerospace Science and Engineering Board and the Board on Army Science and Technology of the National Academies; the AGARD (NATO) advisory panel for aerospace engineering, as President of the American Academy of Mechanics, as Chair of the US National Committee on Theoretical and Applied Mechanics and as Chairman of the National Council of Deans of Engineering. From the AIAA he has received the Structure, Structural Dynamics and Materials Award, the Von Karman Lectureship, the Crichlow Trust Prize and the Reed Aeronautics Award; from the ASME he has received the Spirit of St. Louis Medal, the Den Hartog Award, Lyapunov Medal and the Caughey Medal; and he has also received the Guggenheim Medal which is awarded jointly by the AIAA, ASME, AHS and SAE. He has served on the boards of visitors of several universities and is a consultant to government, industry and universities in science and technology policy and engineering education as well as on the topics of his research. Dr. Dowell research and teaching ranges over the topics of acoustics, aerodynamics, aeroelasticity, dynamics and structures. In addition to being author of over four hundred research articles, Dr. Dowell is the author or co-author of four books, "Aeroelasticity of Plates and Shells", "A Modern Course in Aeroelasticity", "Studies in Nonlinear Aeroelasticity" and “Dynamics of Very High Dimensional Systems”. Dr. Dowell received his B.S. degree from the University of Illinois and his S.M. and Sc.D. degrees from the Massachusetts Institute of Technology. Before coming to Duke as Dean of the School of Engineering, serving from 1983-1999, he taught at M.I.T. and Princeton. He has also worked with the Boeing Company.
Host: Dr. Petros Ioannou, ioannou@usc.edu
More Info: https://forms.gle/zUxvBSDsb1TCHdcEA
Location: Ronald Tutor Hall of Engineering (RTH) - RTH 526
Audiences: Everyone Is Invited
Contact: Miki Arlen
Event Link: https://forms.gle/zUxvBSDsb1TCHdcEA
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, Oct 16, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Alex Robey, Postdoctoral Researcher
Talk Title: Jailbreaking LLM-Controlled Robots
Series: EE598 Seminar Series
Abstract: Recent research has shown that large language models (LLMs) such as OpenAI's ChatGPT are susceptible to jailbreaking attacks, wherein malicious users fool an LLM into generating harmful content (e.g., bombbuilding instructions). However, these attacks are generally limited to eliciting text from chatbots. In contrast, we consider attacks on LLM-controlled robots, which, if jailbroken, could be manipulated into causing physical harm in the real world. Our attacks successfully jailbreak a self-driving LLM, a wheeled Clearpath Robotics Jackal robot, and, most concerningly, the commercially available Unitree Go2 robot dog. In this talk, we will walk through the recent history of jailbreaking, describe our robotic attacks, and discuss how such attacks can be mitigated to avoid the misuse of AI-powered robots.
Biography: Alex Robey is a postdoctoral researcher in the Machine Learning Department at Carnegie Mellon University, where he is advised by J. Zico Kolter. He is also affiliated with Gray Swan, a start-up that aims to develop AI models resistant to adversarial attacks. In 2024, he received his Ph.D. from the Department of Electrical and Systems Engineering at the University of Pennsylvania, where he was advised by Hamed Hassani and George J. Pappas. He was recently named a Rising Star in Adversarial Machine Learning (AdvML) at the NeurIPS 2024 workshop on AdvML, and he was also the recipient of the Best Paper Award from the AdvML workshop at ICML 2023.
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. -
MHI ISSS Seminar - Dr. Ioannis Savidis, Friday, October 18th at 2pm in EEB 132
Fri, Oct 18, 2024 @ 02:00 PM - 03:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Ioannis Savidis, Associate Professor, Drexel University
Talk Title: AI/ML for EDA: Learning Algorithms in Analog and Digital Design
Series: Integrated Systems
Abstract: In the ever-evolving landscape of Electronic Design Automation (EDA), the integration of Artificial Intelligence (AI) and Machine Learning (ML) algorithms with traditional heuristic optimization algorithms has emerged as a transformative force in automated circuit design. This presentation delves into the dynamic intersection of AI/ML and EDA, exploring state-of-the-art techniques shaping the analog and digital physical design space. Machine learning, specifically deep learning, has the potential to significantly improve the accuracy, speed, efficiency, and reliability of EDA tasks such as circuit modeling, simulation, layout design, and optimization. Delving into such cutting-edge advancements, I will describe current AI/ML research performed by the ICE Lab that promises to transcend traditional paradigms, with the goal of enabling designers to navigate complexities with unparalleled efficiency and accuracy. Specifically, a focus on state-of-the-art learning and optimization techniques for the modeling and design of mixed-signal ICs will be presented and discussed. Practical considerations, challenges, and opportunities of ML algorithms for analog and digital circuit design will be discussed, with a focus on the use of such algorithms for prediction and optimization tasks within the EDA design flow.
Biography: Dr. Ioannis Savidis (S'03-M'13-SM'18) is an Associate Professor in the Department of Electrical and Computer Engineering at Drexel University, where he directs the Integrated Circuits and Electronics (ICE) Design and Analysis Laboratory. He received his B.S.E. from Duke University in 2005, and his M.Sc. and Ph.D. from the University of Rochester in 2007 and 2013, respectively. Dr. Savidis has authored over 130 technical papers in peer-reviewed journals and conferences, including a book on Three-Dimensional Integrated Circuit Design and holds 16 issued and five pending patents. His research interests include high-performance digital and mixed-signal integrated circuits, power management for SoC and microprocessor circuits, hardware security, AI/ML algorithms for circuit optimization, and electro-thermal modeling for 2-D and 3-D circuits. Dr. Savidis is a senior member of IEEE and has received two Best Paper Awards, the 2018 NSF CAREER Award, and the 2019 DoD DURIP Award. He serves on organizing committees for several conferences including IEEE HOST, ACM GLSVLSI, and IEEE ISCAS, and on technical program committees for DAC, ICCAD, MLCAD, and others. Dr. Savidis is a member of the VLSI Systems and Applications Technical Committee of the IEEE Circuits and Systems Society and serves on the editorial boards of IEEE Transactions on VLSI Systems, Microelectronics Journal, and ACM Transactions on Design Automation of Electronic Systems.
Host: Hossein Hashemi, Mike Chen and Constantine Sideris
Webcast: https://usc.zoom.us/j/94304141343More Information: MHI_Seminar_Flyer_Savidis_Oct18_2024.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
WebCast Link: https://usc.zoom.us/j/94304141343
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: Murat Arcak
Mon, Oct 21, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Murat Arcak, Professor of Electrical Engineering and Computer Sciences | University of California, Berkeley
Talk Title: Data-Driven Approaches for Estimating Reachable Sets in Complex Dynamical Systems
Series: CSC/CommNetS-MHI Seminar Series
Abstract: The computation of reachable sets is essential for characterizing and verifying the behavior of safety-critical systems. However, many practical systems are high-dimensional and analytically intractable, making the exact computation of reachable sets difficult or impossible. We propose a data-driven approach that uses a finite ensemble of sample trajectories to estimate reachable sets with probabilistic accuracy guarantees. This method is broadly applicable and computationally advantageous, as the main cost comes from simulating a predetermined number of trajectories, which can be parallelized to reduce computation time. We first present a method that uses scenario optimization to construct reachable set estimates as approximate solutions to chance-constrained optimization problems. Next, we use a class of polynomials derived from empirical moment matrices, whose sublevel sets act as nonconvex estimates of the reachable set. These data-driven methods offer scalable solutions for estimating reachable sets in systems with complex dynamics.
Biography: Murat Arcak is a professor at the University of California, Berkeley, where he holds the Robert M. Saunders Endowed Chair. He has a primary appointment in Electrical Engineering and Computer Sciences, and a courtesy appointment in Mechanical Engineering. He earned his B.S. degree in Electrical Engineering from BoÄaziçi University, Istanbul, Turkey, in 1996, and his M.S. and Ph.D. degrees from the University of California, Santa Barbara, in 1997 and 2000. His research focuses on dynamical systems and control theory, with applications in multi-agent systems and transportation. He received a CAREER Award from the National Science Foundation in 2003, the Donald P. Eckman Award from the American Automatic Control Council in 2006, the Control and Systems Theory Prize from the Society for Industrial and Applied Mathematics (SIAM) in 2007, and the Antonio Ruberti Young Researcher Prize from the IEEE Control Systems Society in 2014. He is a member of ACM and SIAM, and a fellow of both IEEE and the International Federation of Automatic Control (IFAC)
Host: Dr. Lars Lindemann, llindema@usc.edu
More Information: 2024.10.21 CSC Seminar - Murat Arcak.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - EEB 248
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. -
ECE Seminar: A plug-and-play acceleration framework for generative AI models on the edge
Tue, Oct 22, 2024 @ 10:00 AM - 11:00 AM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Yanzhi Wang, Associate Professor and Faculty Fellow, Dept. of ECE, Northeastern University
Talk Title: A plug-and-play acceleration framework for generative AI models on the edge
Abstract: In the generative AI era, general users need to apply different base models, fine tuned checkpoints, and LoRAs. Also the data privacy and real-time requirement will favor on-device, local deployment of large-scale generative AI models. It is desirable to develop a "plug-and-play" framework such that users can download any generative AI model, click and run on their own device. This poses significant challenge to the current AI deployment frameworks, which are typically time-consuming and requires human expertise of hardware and code generation. We present our effort of OminiX, which is a first step towards unified library and acceleration of generative AI models across various hardware platforms. Integrating our unique front-end library and back-end instantaneous acceleration techniques, which will be open-source soon, we show capability of plug-and-play deployment and state-of-the-art acceleration of various generative AI models, starting from image generation, large language models, multi-model language models, speech generation and voice cloning, real-time chatting engine, real-time translation, video generation, real-time avatar, to name a few. This can be achieved on everyone's own platform.
Biography: Yanzhi Wang is Associate Professor in the Department of Electrical and Computer Engineering at Northeastern University, a senior member of IEEE. His research interests focus on real-time and energy-efficient deep learning and artificial intelligence systems, especially on efficient large language models and large-scale generative AI systems. His research works have been published broadly in (i) machine learning conferences such as AAAI, CVPR, NeurIPS, ICML, ICCV, ICLR, IJCAI, ECCV, KDD, ICRA, ACM MM, ICDM, etc., (ii) architecture and system conferences such as ASPLOS, ISCA, MICRO, HPCA, CCS, VLDB, PLDI, WWW, ICS, PACT, CGO, IPDPS, INFOCOM, ICDCS, DAC, ICCAD, FPGA, FCCM, ISSCC, CICC, RTAS, RTSS, etc., and (iii) IEEE and ACM transactions. His research works have been cited over 20,500 times. He has received six Best Paper Awards and another 12 Best Paper Nominations. He has received the U.S. Army Research Office Young Investigator Program Award (YIP), IEEE TC-SDM Early Career Award, Asia Pacific Signal and Information Processing Association Distinguished Leader Award, Massachusetts Acorn Innovation Award, design contest awards from multiple conferences, and other research awards from Google, MathWorks, etc. His research work has been reported and cited by around 500 media. He has 13 academic descendants as tenure-track faculty members at University of Minnesota, Michigan State University, University of Georgia, Clemson University, etc.
Host: Dr. Sandeep Gupta, sandeep@usc.edu
Webcast: https://usc.zoom.us/j/98817797740?pwd=OfzLgQ5S1Gbb7b7mxxXe9FgST9u99L.1 (USC NetID Login Required)Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
WebCast Link: https://usc.zoom.us/j/98817797740?pwd=OfzLgQ5S1Gbb7b7mxxXe9FgST9u99L.1 (USC NetID Login Required)
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. -
ECE Pioneer Series: Ming Hsieh
Wed, Oct 23, 2024 @ 03:30 PM - 05:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Ming Hsieh, Entrepreneur and Philanthropist
Talk Title: From Rural Northeastern China to the American National Academy of Engineering: The Transformative Journey of a Young Boy Empowered by Interdisciplinary Electrophysics & Electrical Systems Education
Series: ECE Pioneer Series
Abstract: This seminar traces the transformative journey of Ming Hsieh, from his early education in rural Northeastern China to becoming a distinguished member of the American National Academy of Engineering. Hsieh’s passion for engineering began at the age of 14 in 1970, deeply influenced by his father, an electrical engineer and scientist, who was dedicated to bringing electricity to rural villages in Northern China. Beginning his studies in Semiconductor Physics/Devices at Southern China University of Technology in 1978, Hsieh shifted his academic focus after transferring to the University of Southern California (USC) in 1981. Under the mentorship of Professor Kurt Lehovec in the Electrical Engineering (EE) Electrophysics group, Hsieh integrated the disciplines of electrophysics and electrical systems. This interdisciplinary education laid the foundation for his professional career, where he pioneered innovations at the intersection of physics and systems. His work, ranging from deploying large-scale biometric systems to accelerating breakthroughs in biomedical research and cancer therapeutics, demonstrates the transformative power of a unified approach to engineering and science. Through his story, Hsieh exemplifies how a robust, interdisciplinary education can lead to groundbreaking advancements with a global impact.
Biography: Ming Hsieh, BSEE ’83, MSEE ’84, is co-founder, president, CEO and chairman of the board of Cogent, Inc., one of the top providers of fingerprint identification systems in the United States. His generous gift of $35 million was the largest ever to name an engineering department in the United States. Ten years later, his endowment continues to set the course for electrical engineering’s expansion into new realms of human invention. As our field grows, so too does the quality of our academic standards and the ability of our graduates to meet the challenges of today's global community.
Below is a sampling of stories and events related to Dr. Hsieh:
Greatest Hits, Vol. 1 (2006-2016): The Ming Hsieh Department of Electrical and Computer Engineering - Read about some of our department's greatest achievements in the decade since Ming Hsieh's gift. https://magazine.viterbi.usc.edu/fall-2016/whats-next/greatest-hits-vol-1-2006-2016-the-ming-hsieh-department-of-electrical-engineering/
Q+A: Ming Hsieh - On the 10th anniversary of his naming gift to the USC Viterbi Ming Hsieh Department of Electrical and Computer Engineering, Ming Hsieh spoke about the department, cancer research and his desire to give back. https://magazine.viterbi.usc.edu/fall-2016/whats-next/qa-ming-hsieh/
Putting a Fingerprint on Electrical Engineering - Cogent, Inc. co-founder and Viterbi School alumnus Ming Hsieh has given $35 million to name the School's oldest and most prominent department. Read this cover story about him from the Fall/Winter 2006 USC Viterbi Engineer magazine. https://viterbi.usc.edu/news/news/2006/putting-a-fingerprint.htm
Viterbi School Celebrates a Momentous Gift to Name the Department of Electrical and Computer Engineering - The Viterbi School celebrates a record-breaking gift and naming of the third department in its history at a red-carpet ceremony held in the Ronald Tutor Hall Courtyard. https://viterbi.usc.edu/news/news/2006/usc-alumnus-ming.htm
Host: Dr. Richard Leahy, leahy@usc.edu
More Information: Screenshot 2024-10-23 at 10.59.36 AM.png
Location: Hughes Aircraft Electrical Engineering Center (EEB) - EEB 132
Audiences: Everyone Is Invited
Contact: Cathy Huang
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. -
ISSS - Dr. Hyun-Sik Kim, Friday, October 25th at 2pm in EEB 132 and Zoom
Fri, Oct 25, 2024 @ 02:00 PM - 03:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Hyun-Sik Kim, Associate Professor, KAIST
Talk Title: Exploring Ways to Maximize Efficiency and Performance in Low-Dropout (LDO) Regulators
Series: Integrated Systems
Abstract: Low-dropout (LDO) regulators are ideal off- and on-chip solutions for powering noise-sensitive loads due to their ripple-less output. LDOs also have many benefits over switch-mode dc-dc converters, such as rapid transient response, excellent power supply rejection (PSR), and compact footprint. Unfortunately, they suffer from an inescapable disadvantage: poor power efficiency; this is primarily caused by a considerable dropout voltage (VDO). Reducing VDO to improve efficiency often leads to a significant drop in LDO's regulation performance. Because of this, most LDOs are designed with a large VDO, making them perceived as energy-consuming components of power management systems. This talk will delve into effective ways to extremely minimize the dropout voltage without compromising performance, aiming for energy-efficient LDO regulators. We will begin with a thorough investigation of operational principles, analyses, and strategies, exploring trade-offs among key performance metrics. Next, several promising approaches to realizing energy-efficient LDO regulators will be investigated, including traditional digital LDOs, a dual-rail analog/digital-hybrid LDO, a triode-region LDO, and a voltage/current-hybrid (VIH) LDO. Finally, the technical merits and flaws of each high-efficiency LDO topology will be investigated by comparing them. In this talk, I will also share my insights from my experience developing the VIH LDO regulator that achieves 98.6% efficiency and a -75dB PSR at 30kHz.
Biography: Hyun-Sik Kim is currently an Associate Professor of Electrical Engineering at the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea. He received his B.S. degree (Hons.) in electronic engineering from Hanyang University, Seoul, South Korea, in 2009, and his M.S. and Ph.D. degrees in electrical engineering from KAIST, in 2011 and 2014, respectively. His research interests include the CMOS analog-integrated circuit designs, with an emphasis on display drivers, power managements, and sensory readout chips. Prof. Kim was a recipient of two Gold Prizes in the 18th and 19th Samsung Human-Tech Paper Awards in 2012 and 2013, respectively, the IEEE SSCS Pre-Doctoral Achievement Award in 2014, the IEEE SSCS Seoul Chapter Best Student JSSC Paper Award in 2014, and the KAIST Technology Innovation Award in 2022. He served as a Guest Editor for the IEEE SOLID-STATE CIRCUITS LETTERS (SSC-L) and was a member of the Technical Program Committee (TPC) for the IEEE Asian Solid-State Circuits Conference (A-SSCC) from 2016 to 2023. He is currently serving on the TPC for the IEEE International Solid-State Circuits Conference (ISSCC) and is the TPC Subcommittee Chair for the IEEE Custom Integrated Circuits Conference (CICC). He has been appointed as a Distinguished Lecturer (DL) in the IEEE Solid-State Circuits Society (SSCS) for the term 2024-2026.
Host: MHI - ISSS, Hashemi, Chen and Sideris
More Info: https://usc.zoom.us/j/92995535728
More Information: MHI_Seminar_Flyer_Kim_Oct25_2024.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
Audiences: Everyone Is Invited
Contact: Marilyn Poplawski
Event Link: https://usc.zoom.us/j/92995535728
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: Jorge Cortes
Mon, Oct 28, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Jorge Cortes, Professor and Cymer Corporation Endowed Chair in High Performance Dynamic Systems Modeling and Control | Department of Mechanical and Aerospace Engineering, University of California, San Diego
Talk Title: The safe gradient flow: a system-theoretic approach to anytime constrained optimization through control barrier functions
Series: CSC/CommNetS-MHI Seminar Series
Abstract: Problems where the solution to a constrained optimization problem is used to regulate a physical process modeled as a dynamically evolving plant arise in multiple application areas, including safety-critical control, power networks, traffic networks, and network congestion. This may take the form of providing setpoints, specifying optimization-based controllers, or steering the system toward an optimal steady state. A paradigmatic example is the use of CBF-based quadratic programs for controller synthesis in robotics. In this talk, we are motivated by situations where the problem incorporates constraints which, when violated, threaten the safe operation of the physical system. In such cases, the algorithm that solves the optimization must be anytime, meaning that it is guaranteed to return a feasible point even when terminated before it has converged to a solution. We introduce a class of novel system-theoretic algorithms for solving constrained nonlinear programs that combine continuous-time gradient flows to optimize the objective function with techniques from control barrier functions to maintain forward invariance of the feasible set. We refer to the resulting closed-loop system as the safe gradient flow. We draw on the alternative interpretation of the safe gradient flow as a projected dynamical system to characterize its dynamical properties regarding regularity, stability, convergence, contractivity, and invariance. We also show how the proposed framework is conducive to the extension of the proposed designs to monotone variational inequalities and discrete-time settings.
Biography: Jorge Cortes is a Professor and Cymer Corporation Endowed Chair in High Performance Dynamic Systems Modeling and Control in the Department of Mechanical and Aerospace Engineering, University of California, San Diego. He is the author of "Geometric, Control and Numerical Aspects of Nonholonomic Systems" (New York: Springer-Verlag, 2002) and co-author of "Distributed Control of Robotic Networks” (Princeton: Princeton University Press, 2009). He is a Fellow of IEEE, SIAM, and IFAC. He has co-authored papers that have won the 2008 and the 2021 IEEE Control Systems Outstanding Paper Award, the 2009 SIAM Review SIGEST selection from the SIAM Journal on Control and Optimization, the 2012 O. Hugo Schuck Best Paper Award in the Theory category, and the 2019 and 2023 IEEE Transactions on Control of Network Systems Outstanding Paper Award. At the IEEE Control Systems Society, he has been a Distinguished Lecturer (2010-2014), an elected member (2018-2020) of the Board of Governors, and Director of Operations (2019-2022) of its Executive Committee. His research interests include distributed control and optimization, network science and complex systems, resource-aware control and coordination, distributed decision making and autonomy, network neuroscience, and multi-agent coordination in robotic, power, and transportation networks.
Host: Dr. Lars Lindemann, llindema@usc.edu
More Information: 2024.10.28 CSC Seminar - Jorge Cortes.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. -
IEEE-GRSS-APS-SSCS Joint seminar - Stefano Maci, Wed. Oct. 30th at 10am in RTH 211 and Zoom
Wed, Oct 30, 2024 @ 10:00 AM - 11:00 AM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Prof. Stefano Maci, IEEE AP-S Past President, University of Sienna, Italy
Talk Title: Metasurface Antennas
Series: IEEE GRSS-APS-SSCS
Abstract: Metasurfaces belong to the category of thin metamaterials and find applications across a wide frequency range, from microwaves to optical frequencies, for developing innovative electromagnetic engineering devices. These surfaces are created by densely arranging small elements on or etching them into a dielectric substrate in a locally periodic distribution. By adjusting the dimensions of these elements while maintaining sub-wavelength 2D periodicity, a pixelated visual appearance and an electromagnetic modulation of the equivalent local impedance boundary conditions (IBC) are achieved. The manipulation of IBC allows for localized modifications in the dispersion equation, influencing the local wavevector while maintaining a constant operating frequency. This capability enables the transformation of surface or guided waves into various wavefield configurations with specified properties. This presentation will focus on the control of both surface waves and space waves, showcasing examples such as the design of high-gain, low cross-polarization antennas, multibeam antennas, and scanning beam flat lenses. Emphasis will be given to space applications. The discussion will also delve into the third generation of adaptive metasurfaces (MTSs), featuring dynamically reconfigurable boundary conditions. This advancement opens possibilities for exploring new perspectives in the development of next-generation wireless communication systems.
Biography: Prof. Stefano Maci is a Professor at the University of Siena (UNISI). Since 2000, he has been P.I. of 10 research projects funded by the European Union (EU) and by the European Space Agency (ESA). He is a Fellow of IEEE since 2004. In 2004 he founded the European School of Antennas (ESoA), a PhD school that presently comprises 35 courses on Antennas, Propagation, and Electromagnetic Theory, and 200 teachers, among them 20 IEEE Fellow. He has been advisor of 40 PhD students. He has been former member of IEEE Antennas and Propagation Society (AP-S) AdCom, the Chair of the Award Committee of the IEEE AP-S, member of the AP Executive Board of IET (UK), Distinguished Lecturer of IEEE and of EurAAP. He was recipient of several prizes and awards, among which the EurAAP Award 2014, the Chen-To Tai Distinguished Educator award 2016, of the Shelkunoff Transaction Prize in 2015, and of the URSI Dellinger Gold Medal in 2020. He is presently Director of ESoA. He has been TPC Chair of the METAMATERIAL 2020 and and General Chair of EuCAP 2023. He was the president of the IEEE Antennas and Propagation Society 2023. In the last ten years he has been invited 60 times as key-note speaker in international conferences. His research activity is documented in 200 papers published in international journals, (among which 100 on IEEE journals), 10 book chapters, and about 450 papers in proceedings of international conferences.
Host: IEEE GRSS-APS-SSCS Joint Student Chapter
More Info: Meeting ID: 925 1030 8883, Passcode: 613281
More Information: IEEE Stefano Maci.pdf
Location: Ronald Tutor Hall of Engineering (RTH) - 211
Audiences: Everyone Is Invited
Contact: Marilyn Poplawski
Event Link: Meeting ID: 925 1030 8883, Passcode: 613281
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, Oct 30, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Anushri Dixit, Assistant Professor in the Department of Mechanical & Aerospace Engineering at the University of California, Los Angeles
Talk Title: Perceive with Confidence: Statistical Safety Assurances for Vision-based Navigation
Series: EE598 Seminar Series
Abstract: Significant strides in perception over the past few years have enabled robotic systems to interpret and interact with the world in increasingly versatile ways. The large, often multi-modal, datasets that are used to train modern perception systems endow robots with capabilities for scene understanding like object detection and segmentation. However, the safe integration and reliability of these learned perception models for robotic applications still remains in question due to their failures in unfamiliar environments. In this talk, I will discuss our framework, Perceive with Confidence (PwC), for rigorously quantifying the uncertainty of a pre-trained obstacle detection system in a way that provides a formal assurance on correctness and safety for planning applications. This is achieved by utilizing a technique called conformal prediction to calibrate the perceptual outputs while ensuring generalization to novel environments. I will provide experimental validations of PwC’s formal assurances for indoor navigation applications on the Unitree Go1 quadruped.
Biography: Anushri Dixit is an Assistant Professor in the Department of Mechanical & Aerospace Engineering at the University of California, Los Angeles. Prior to UCLA, she was a postdoctoral researcher at Princeton University. She received her Ph.D. in Control and Dynamical Systems from California Institute of Technology in 2023 and her B.S. in Electrical Engineering from Georgia Institute of Technology in 2017. Her research focuses on motion planning and control of robots in unstructured environments while accounting for uncertainty in a principled manner. Her work on risk-aware methodologies for planning has been deployed on various robotic platforms as a part of Team CoSTAR’s effort in the DARPA Subterranean Challenge. She has received the Outstanding Student Paper Award at the Conference on Decision and Control, Best Student Paper Award at the Conference of Robot Learning, and was selected as a Rising Star in Data Science by The University of Chicago.
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. -
AIF4S Seminar: Value of Pretraining Data: Scaling Laws for Downstream Task Performance of Large Language Models
Thu, Oct 31, 2024 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Berivan Isik, Research Scientist, Google, Inc.
Talk Title: Value of Pretraining Data: Scaling Laws for Downstream Task Performance of Large Language Models
Abstract: This talk explores the challenges and open questions surrounding the value of pretraining data for large language models (LLMs) in transfer learning settings. While scaling laws have provided valuable insights for LLM design, existing work has predominantly focused on pretraining loss. In contrast, this work investigates scaling behavior in a transfer learning setting where LLMs are finetuned for downstream tasks. Specifically, we examine how the choice and size of pretraining data impact downstream performance, as measured by cross-entropy and translation quality metrics such as BLEU and COMET. Our experiments reveal that the size of the finetuning dataset and the alignment between pretraining and downstream data significantly influence scaling behavior. With sufficient alignment, both cross-entropy and translation quality improve with increased pretraining data, and we demonstrate the ability to predict translation quality using a new log-law. However, in cases of moderate misalignment, we observe that translation quality can fluctuate or even deteriorate with more pretraining data, despite consistent improvements in cross-entropy. Through analysis of these findings, we provide insights for selecting appropriate pretraining data. The talk will conclude with a discussion of future research directions and remaining open questions in this area.
Biography: Berivan Isik is a research scientist at Google, working on efficient and trustworthy AI. Her current interests are efficient training/finetuning of large models, pretraining data valuation and scaling laws for LLMs, differential privacy, and unlearning. She earned her PhD from Stanford University in 2024, where she was affiliated with the SAIL and StatsML groups. Her research was supported by Stanford Graduate Fellowship (2019-2023), Google Ph.D. Fellowship (2023-2026), and a Meta research grant.
Host: Dr. Mahdi Soltanolkotbi, soltanol@usc.edu
Webcast: https://usc.zoom.us/j/98648507063?pwd=kORhNLFVMLol7FYlHv6TsAmqcKqD7t.1Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
WebCast Link: https://usc.zoom.us/j/98648507063?pwd=kORhNLFVMLol7FYlHv6TsAmqcKqD7t.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.
