BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20230312T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20221106T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20230312T201719Z UID:00D00372-61D2-484D-BBC9-DEA29CCCE245 DTSTART;TZID=America/New_York:20230309T120000 DTEND;TZID=America/New_York:20230309T130000 DESCRIPTION:Seminar: Lifelong Learning – Temporal Aspects and Theoretical Foundations\n\nSpeaher: Hava Siegelmann\, PhD\nProvost Professor of Compu ter Science\,\nDirector of the Biologically Inspired\nNeural and Dynamical Systems (BINDS) Laboratory\,\nUniversity of Massachusetts Amherst\n\nTime /date: Thursday\, March 9th\n12:00-1:00 pm EST\nGeorge Mason Fairfax Campu s:\nHorizon Hall\, Rm 2008\nLive streaming to SciTech Campus: KJH 258\n\nA bstract\nLifelong Learning is the cutting edge of artificial intelligence - encompassing\ncomputational methods that allow systems to learn in runti me and incorporate\nlearning for application in new\, unanticipated situat ions. Until recently\, this sort of\ncomputation has been found exclusivel y in nature\; thus\, Lifelong Learning looks\nto nature for its underlying principles and mechanisms and then transfer them to\nthis new technology. Yet SOTA Lifelong Learning\, like classical learning\, is limited\nin its accuracy for temporal prediction from short and incomplete measurements.\ nThis is where our new technology appears. The technology is based on a ne w\ntype of neural networks\, where much like the brain\, the connections b etween\nneurons are no longer scalar numbers\, but rather temporal functio ns. This gives\nthe networks an unparallel capacity\, strong temporal accu racy\, and ability to\nkeep effectivity even when most measurements are lo st. Interestingly\, our\ntemporally changing network while more capable\, is smaller in size and\nconsumes significantly less power. A version used for Reinforcement learning and\ncontrol is under development. We will also introduce the forward propagation\nalgorithm which my lab pioneers. Compu tational foundations are required to\nenable new computation to evolve. Wh ile Turing computation worked for us till\nnow very well\, it does not wel l describe such machines that change regularly.\nIndeed Turing spent most of his research time\, immediately after introducing the\nlogical universa l model\, in an effort to find a new model that will more closely\nsimulat e the brains and will learn more like them. Super-Turing computation may\n be an answer to Turing’s quest\, and since it is built on every learning machine\, it\nhas become the foundational theory of Lifelong Learning\, t owards a stronger AI.\n\nBiography\nDr. Siegelmann is a professor of Compu ter Science\, Core Member of the\nNeuroscience and Behavior Program\, and director of the Biologically Inspired\nNeural and Dynamical Systems (BINDS ) Laboratory. Siegelmann recently\ncompleted her term as a DARPA PM: “L2 M\,” one of her key initiatives\,\ninaugurated “third-wave AI\,” pus hing major design innovation and a dramatic\nincrease in AI capability. “GARD” is leading to unique advancements in assuring\nAI robustness ag ainst attack. “CSL” is introducing powerful methods of combined\nlearn ing and information sharing on AI platforms without revealing private data .\nOther programs include advanced biomedical applications. Siegelmann\nco nducts highly interdisciplinary research in next generation machine learni ng\,\nneural networks\, intelligent machine-human collaboration\, computat ional\nstudies of the brain - with application to AI\, data science and\ni ndustrial/government /biomedical applications. Among her contributions are \nthe Support Vector Clustering algorithm\, delineating jet-lag mechanisms \,\nidentifying brain structure that leads to abstract thoughts\, and Supe r-Turing\ntheory which has become the backbone of the latest generation of biologically\ninspired neural networks and lifelong learning machines. Dr . Siegelmann is a\nleader in increasing awareness of ethical AI via the IE EE\, INNS and international\nmeetings\, and is particularly active in supp orting minorities and women in STEM\nnationally and internationally curren tly serving as the Chair of the women’s\nchapter of the International Ne ural Networks Society. Siegelmann has been a\nvisiting professor at MIT\, Harvard University\, the Weizmann Institute\, ETH\, the Salk\nInstitute\, Mathematical Science Research Institute Berkeley\, and the Newton\nInstitu te Cambridge University. She is the former PM for L2M\, DARPA’s largest\ nadvanced AI initiative\, as well as other major DARPA programs. She was t he\nrecipient of the Alon Fellowship of Excellence\, the NSF-NIH Obama Pre sidential\nBRAIN Initiative award\, the Donald O. Hebb Award of the Intern ational Neural\nNetwork Society for “contribution to biological learning ”\; she was named named\nIEEE fellow and Distinguished Lecturer of the I EEE Computational Intelligence\nSociety\, and INNS fellow. She received th e DARPA Meritorious Public Service\naward.\n\nRoom: 2008\, Bldg: Horizon H all\, George Mason University - Fairfax campus\, Fairfax\, Virginia\, Unit ed States\, 22030 LOCATION:Room: 2008\, Bldg: Horizon Hall\, George Mason University - Fairfa x campus\, Fairfax\, Virginia\, United States\, 22030 ORGANIZER:kafi@ieee.org SEQUENCE:3 SUMMARY:Lifelong Learning – Temporal Aspects and Theoretical Foundations URL;VALUE=URI:https://events.vtools.ieee.org/m/351372 X-ALT-DESC:Description: <br /><p><strong>Seminar: Lifelong Learning &ndash\ ; Temporal Aspects and Theoretical Foundations</strong></p>\n<p><strong>Sp eaher: Hava Siegelmann\, PhD</strong><br />Provost Professor of Computer S cience\,<br />Director of the Biologically Inspired<br />Neural and Dynami cal Systems (BINDS) Laboratory\,<br />University of Massachusetts Amherst< /p>\n<p><strong>Time/date: </strong><strong>Thursday\, March 9th<br />&nbs p\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\ ; 12:00-1:00 pm EST<br />&nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; & nbsp\; &nbsp\; &nbsp\; &nbsp\; George Mason Fairfax Campus:<br />&nbsp\; & nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; Hor izon Hall\, Rm 2008<br />&nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; & nbsp\; &nbsp\; &nbsp\; &nbsp\; Live streaming to SciTech Campus: KJH 258&n bsp\;</strong></p>\n<p><strong>Abstract</strong><br />Lifelong Learning is the cutting edge of artificial intelligence - encompassing<br />computati onal methods that allow systems to learn in runtime and incorporate<br />l earning for application in new\, unanticipated situations. Until recently\ , this sort of<br />computation has been found exclusively in nature\; thu s\, Lifelong Learning looks<br />to nature for its underlying principles a nd mechanisms and then transfer them to<br />this new technology. Yet SOTA Lifelong Learning\, like classical learning\, is limited<br />in its accu racy for temporal prediction from short and incomplete measurements.<br /> This is where our new technology appears. The technology is based on a new <br />type of neural networks\, where much like the brain\, the connection s between<br />neurons are no longer scalar numbers\, but rather temporal functions. This gives<br />the networks an unparallel capacity\, strong te mporal accuracy\, and ability to<br />keep effectivity even when most meas urements are lost. Interestingly\, our<br />temporally changing network wh ile more capable\, is smaller in size and<br />consumes significantly less power. A version used for Reinforcement learning and<br />control is unde r development. We will also introduce the forward propagation<br />algorit hm which my lab pioneers. Computational foundations are required to<br />e nable new computation to evolve. While Turing computation worked for us ti ll<br />now very well\, it does not well describe such machines that chang e regularly.<br />Indeed Turing spent most of his research time\, immediat ely after introducing the<br />logical universal model\, in an effort to f ind a new model that will more closely<br />simulate the brains and will l earn more like them. Super-Turing computation may<br />be an answer to Tur ing&rsquo\;s quest\, and since it is built on every learning machine\, it< br />has become the foundational theory of Lifelong Learning\, towards a s tronger AI.</p>\n<p><strong>Biography</strong><br />Dr. Siegelmann is a pr ofessor of Computer Science\, Core Member of the<br />Neuroscience and Beh avior Program\, and director of the Biologically Inspired<br />Neural and Dynamical Systems (BINDS) Laboratory. Siegelmann recently<br />completed h er term as a DARPA PM: &ldquo\;L2M\,&rdquo\; one of her key initiatives\,< br />inaugurated &ldquo\;third-wave AI\,&rdquo\; pushing major design inno vation and a dramatic<br />increase in AI capability. &ldquo\;GARD&rdquo\; is leading to unique advancements in assuring<br />AI robustness against attack. &ldquo\;CSL&rdquo\; is introducing powerful methods of combined<br />learning and information sharing on AI platforms without revealing priv ate data.<br />Other programs include advanced biomedical applications. Si egelmann<br />conducts highly interdisciplinary research in next generatio n machine learning\,<br />neural networks\, intelligent machine-human coll aboration\, computational<br />studies of the brain - with application to AI\, data science and<br />industrial/government /biomedical applications. Among her contributions are<br />the Support Vector Clustering algorithm\ , delineating jet-lag mechanisms\,<br />identifying brain structure that l eads to abstract thoughts\, and Super-Turing<br />theory which has become the backbone of the latest generation of biologically<br />inspired neural networks and lifelong learning machines. Dr. Siegelmann is a<br />leader in increasing awareness of ethical AI via the IEEE\, INNS and internationa l<br />meetings\, and is particularly active in supporting minorities and women in STEM<br />nationally and internationally currently serving as the Chair of the women&rsquo\;s<br />chapter of the International Neural Netw orks Society. Siegelmann has been a<br />visiting professor at MIT\, Harva rd University\, the Weizmann Institute\, ETH\, the Salk<br />Institute\, M athematical Science Research Institute Berkeley\, and the Newton<br />Inst itute Cambridge University. She is the former PM for L2M\, DARPA&rsquo\;s largest<br />advanced AI initiative\, as well as other major DARPA program s. She was the<br />recipient of the Alon Fellowship of Excellence\, the N SF-NIH Obama Presidential<br />BRAIN Initiative award\, the Donald O. Hebb Award of the International Neural<br />Network Society for &ldquo\;contri bution to biological learning&rdquo\;\; she was named named<br />IEEE fell ow and Distinguished Lecturer of the IEEE Computational Intelligence<br /> Society\, and INNS fellow. She received the DARPA Meritorious Public Servi ce<br />award.&nbsp\;&nbsp\;</p>\n<p>&nbsp\;</p> END:VEVENT END:VCALENDAR