BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20261101T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260916T151607Z UID:51F79BF9-B08D-4049-8837-4B57C48CBB9D DTSTART;TZID=America/New_York:20261002T153000 DTEND;TZID=America/New_York:20261002T163000 DESCRIPTION:IEEE CASS Distinguished Lecture\n\nSpeaker: Dr. Abhronil Sengup ta\n\nTitle: Opportunities in NeuroAI: A Hardware-Software Co-Design Persp ective\nAbstract: The massive explosion of foundation models in the past f ew years has led to widespread\nadoption of AI in a plethora of applicatio ns. However\, the computational resource requirements to enable\nthese pla tforms are growing at an unsustainable rate. Further\, deep learning syste ms also suffer from\nalgorithmic challenges such as the requirement of a h uge amount of labelled data\, adversarial\nsusceptibility\, lack of interp retability\, etc. Brain-inspired neuromorphic computing has the potential to\novercome these concerns of current AI systems. This talk reviews recen t developments in the domain of\nNeuroAI from my group guided by an overar ching system-science perspective. From the top-down\nalgorithm side\, I wi ll delve into methodologies that treat neuromorphic architectures as conti nuously\nevolving dynamical systems\, revealing intriguing parallels with the learning dynamics in the brain. Our\ndeveloped methodologies enable br ain-inspired models to transition beyond simple vision-related tasks to\nc omplex sequence learning problems trained using online local learning. Com plementary to this effort\, I\nwill also elaborate on a bottom-up perspect ive of bridging the computational efficiency gap through the\nexploration of novel ferroelectric devices that provide a better match to the computat ional primitives of AI\nsystems. Our proposed Compute-In-Memory neuromorph ic processors enabled with in-situ learning\nfunctionalities can potential ly achieve two to three orders of magnitude energy improvement compared to \nstate-of-the-art CMOS implementations. From a system implementation stan dpoint\, I will outline several\nhardware-software co-design strategies to enable variation-aware\, robust\, self-healing AI hardware. I will\nconcl ude my talk with my vision of expanding the scope of neuromorphic computin g beyond simple\nneurons and synapses by forging stronger connections with computational neuroscience\, thereby\nenabling a new generation of brain- inspired computers.\nSpeaker Bio: Dr. Abhronil Sengupta is an Associate Pr ofessor in the School of Electrical Engineering\nand Computer Science at P enn State University. Dr. Sengupta received the PhD degree in Electrical a nd\nComputer Engineering from Purdue University in 2018 and the B.E. degre e from Jadavpur University\,\nIndia in 2013. He worked as a DAAD (German A cademic Exchange Service) Fellow at the University of\nHamburg\, Germany i n 2012\, and as a graduate research intern at Intel Labs in 2016 and Meta Reality\nLabs in 2017.\nDr. Sengupta is pursuing an inter-disciplinary res earch agenda at the intersection of hardware and\nsoftware across the stac k of sensors\, devices\, circuits\, systems and algorithms for enabling ad aptive\,\nscalable and low-power brain-inspired computing systems. He has published over 100 articles in refereed\njournals and conferences and hold s 3 US patents. He has received multiple prestigious recognitions\,\ninclu ding the Penn State Engineering Alumni Society (PSEAS) Outstanding Researc h Award (2026)\,\ninvited participant in NAE Grainger Foundation Frontiers of Engineering Symposium (2026)\, IEEE CEDA\nErnest Kuh Early Career Awar d (2025)\, ARO Early Career Award (2024)\, Purdue Engineering 38 by 38\nAw ard (2024)\, NSF CAREER Award (2023)\, IEEE EDS Chenming Hu Early Career A ward (2023)\,\nMonkowski Career Development Professorship from Penn State University (2022-2025)\, IEEE CASS\nOutstanding Young Author Award (2019)\ , and Meta Faculty Award (2018). His work on neuromorphic\ncomputing has b een highlighted in media by MIT Technology Review\, US Department of Defen se\,\nAmerican Institute of Physics\, IEEE Spectrum\, and Nature Materials . Dr. Sengupta is a Senior Member of\nthe NAI\, IEEE and ACM. He currently serves as an ACM Distinguished Speaker (2024-2027) and IEEE\nCASS Disting uished Lecturer (2025-2026). He is also a co-author of the book &quot\;Int roduction to\nNeuromorphic Computing&quot\; published by Cambridge Univers ity Press.\n\nLocation: 8228 Paint Branch Dr\, College Park\, MD 20742\, Z upnik Lecture Hall 1110\, Kim Building\, University of Maryland\, College Park\n\nSpeaker(s): Dr. Abhronil Sengupta\, \n\nAgenda: \nSeminar by Dr. A bhronil Sengupta\n\nRoom: 1110\, Bldg: Kim Building\, 8228 Paint Branch Dr \, College Park\, Maryland\, United States\, 20742 LOCATION:Room: 1110\, Bldg: Kim Building\, 8228 Paint Branch Dr\, College P ark\, Maryland\, United States\, 20742 ORGANIZER:sshah389@umd.edu SEQUENCE:24 SUMMARY:IEEE CASS Distinguished Lecture URL;VALUE=URI:https://events.vtools.ieee.org/m/577159 X-ALT-DESC:Description: <br /><p>IEEE CASS Distinguished Lecture</p>\n<p>Sp eaker: Dr. Abhronil Sengupta</p>\n<p>Title: Opportunities in NeuroAI: A Ha rdware-Software Co-Design Perspective<br>Abstract: The massive explosion o f foundation models in the past few years has led to widespread<br>adoptio n of AI in a plethora of applications. However\, the computational resourc e requirements to enable<br>these platforms are growing at an unsustainabl e rate. Further\, deep learning systems also suffer from<br>algorithmic ch allenges such as the requirement of a huge amount of labelled data\, adver sarial<br>susceptibility\, lack of interpretability\, etc. Brain-inspired neuromorphic computing has the potential to<br>overcome these concerns of current AI systems. This talk reviews recent developments in the domain of <br>NeuroAI from my group guided by an overarching system-science perspect ive. From the top-down<br>algorithm side\, I will delve into methodologies that treat neuromorphic architectures as continuously<br>evolving dynamic al systems\, revealing intriguing parallels with the learning dynamics in the brain. Our<br>developed methodologies enable brain-inspired models to transition beyond simple vision-related tasks to<br>complex sequence learn ing problems trained using online local learning. Complementary to this ef fort\, I<br>will also elaborate on a bottom-up perspective of bridging the computational efficiency gap through the<br>exploration of novel ferroele ctric devices that provide a better match to the computational primitives of AI<br>systems. Our proposed Compute-In-Memory neuromorphic processors e nabled with in-situ learning<br>functionalities can potentially achieve tw o to three orders of magnitude energy improvement compared to<br>state-of- the-art CMOS implementations. From a system implementation standpoint\, I will outline several<br>hardware-software co-design strategies to enable v ariation-aware\, robust\, self-healing AI hardware. I will<br>conclude my talk with my vision of expanding the scope of neuromorphic computing beyon d simple<br>neurons and synapses by forging stronger connections with comp utational neuroscience\, thereby<br>enabling a new generation of brain-ins pired computers.<br>Speaker Bio: Dr. Abhronil Sengupta is an Associate Pro fessor in the School of Electrical Engineering<br>and Computer Science at Penn State University. Dr. Sengupta received the PhD degree in Electrical and<br>Computer Engineering from Purdue University in 2018 and the B.E. de gree from Jadavpur University\,<br>India in 2013. He worked as a DAAD (Ger man Academic Exchange Service) Fellow at the University of<br>Hamburg\, Ge rmany in 2012\, and as a graduate research intern at Intel Labs in 2016 an d Meta Reality<br>Labs in 2017.<br>Dr. Sengupta is pursuing an inter-disci plinary research agenda at the intersection of hardware and<br>software ac ross the stack of sensors\, devices\, circuits\, systems and algorithms fo r enabling adaptive\,<br>scalable and low-power brain-inspired computing s ystems. He has published over 100 articles in refereed<br>journals and con ferences and holds 3 US patents. He has received multiple prestigious reco gnitions\,<br>including the Penn State Engineering Alumni Society (PSEAS) Outstanding Research Award (2026)\,<br>invited participant in NAE Grainger Foundation Frontiers of Engineering Symposium (2026)\, IEEE CEDA<br>Ernes t Kuh Early Career Award (2025)\, ARO Early Career Award (2024)\, Purdue E ngineering 38 by 38<br>Award (2024)\, NSF CAREER Award (2023)\, IEEE EDS C henming Hu Early Career Award (2023)\,<br>Monkowski Career Development Pro fessorship from Penn State University (2022-2025)\, IEEE CASS<br>Outstandi ng Young Author Award (2019)\, and Meta Faculty Award (2018). His work on neuromorphic<br>computing has been highlighted in media by MIT Technology Review\, US Department of Defense\,<br>American Institute of Physics\, IEE E Spectrum\, and Nature Materials. Dr. Sengupta is a Senior Member of<br>t he NAI\, IEEE and ACM. He currently serves as an ACM Distinguished Speaker (2024-2027) and IEEE<br>CASS Distinguished Lecturer (2025-2026). He is al so a co-author of the book &amp\;quot\;Introduction to<br>Neuromorphic Com puting&amp\;quot\; published by Cambridge University Press.<br><br>Locatio n: 8228 Paint Branch Dr\, College Park\, MD 20742\, Zupnik Lecture Hall 11 10\, Kim Building\, University of Maryland\, College Park</p><br /><br />A genda: <br /><p>Seminar by Dr. Abhronil Sengupta</p> END:VEVENT END:VCALENDAR