BEGIN:VCALENDAR VERSION:2.0 PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4// BEGIN:VEVENT UID:20260710T033437EDT-4464LZdrMT@132.216.98.100 DTSTAMP:20260710T073437Z DESCRIPTION:Clustering and generalization of abstract structures in reinfor cement learning and musicality\n\nBy Michael J. Frank\n\nEdgar L. Marston Professor of Cognitive\, Linguistic & Psychological Sciences at Brown Univ ersity.\n\nWith High-Level Panel of Leaders in Science\, Technology\, On-t he-Ground Action\, and Policy\n\nRegister & watch the webinar\n\nHumans ar e remarkably adept at generalizing knowledge between experiences in a way that can be difficult for computers. Previous computational models and dat a suggest that rather than learning about each individual context\, humans build latent abstract structures and learn to link these structures to ar bitrary contexts\, facilitating generalization\, but with a cost in effici ency of initial learning. In these models\, task structures that are more popular across contexts are likely to be reused in new contexts. Neural si gnatures of such structure learning are predictive across individuals of t he ability to transfer knowledge to new situations. However\, these models predict that structures are either re-used as a whole or created from scr atch\, prohibiting the ability to generalize constituent parts of learned structures. This contrasts with ecological settings\, where task structure s can be decomposed into constituent parts and reused in a compositional f ashion. Moreover in many situations people can transfer structures that th ey have learned to entirely new situations\, by analogy\, even when surfac e aspects of the transition and reward functions change. I will present no vel computational models across levels (from neural networks to bayesian f ormulations) that address how agents and humans can learn and generalize s uch abstract and compositional structure. Throughout\, I will give example s of how such computations can allow a musician to learn to compositionall y transfer musical scales and rhythms within and across instruments. Discu ssion with panelists will follow on the similarity/dissimilarity between h uman and machine in such abstraction.\n\nAbout the speaker\n\nMichael J. F rank is Edgar L. Marston Professor of Cognitive\, Linguistic & Psychologic al Sciences at Brown University. He directs the Center for Computational B rain Science within the Carney Institute for Brain Science. He received hi s PhD in Neuroscience and Psychology in 2004 at the University of Colorado \, following undergraduate and master's degrees in electrical engineering. Frank’s work focuses primarily on theoretical models of frontostriatal ci rcuits and their modulation by dopamine\, especially their cognitive funct ions and implications for neurological and psychiatric disorders. The mode ls are tested and refined with experiments across species\, neural recordi ng methods\, and neuromodulation. Honors include the Troland Research Awar d from the National Academy of Sciences (2021)\, Kavli Fellow (2016)\, the Cognitive Neuroscience Society Young Investigator Award (2011)\, and the Janet T Spence Award for early career transformative contributions (Associ ation for Psychological Science\, 2010). Dr Frank is a senior editor for e Life.\n\n\nAbout the series\n\nThe Precision Convergence series is launche d to catalyze unique synergy between\, on the one hand\, novel partnership s across sciences\, sectors and jurisdictions around targeted domains of r eal-world solutions\, and on the other hand\, a next generation convergenc e of AI with advanced research computing and other data and digital archit ectures such as PSC’s Bridges-2\, and supporting data sharing frameworks s uch as HuBMAP\, informing in a real time as possible the design\, deployme nt and monitoring of solutions for adaptive real-world behaviour and conte xt.\n\nThe Precision Convergence Webinar Series is co-hosted by The 91ºÚÁÏÍø Centre for the Convergence of Health and Economics (MCCHE) at 91ºÚÁÏÍø Univ ersity and The Pittsburgh Supercomputing Center\, a joint computational re search centre between Carnegie Mellon University and the University of Pit tsburgh.\n\n \n DTSTART:20221109T160000Z DTEND:20221109T180000Z SUMMARY:MCCHE Precision Convergence Webinar Series with Michael J. Frank URL:/desautels/channels/event/mcche-precision-converge nce-webinar-series-michael-j-frank-343330 END:VEVENT END:VCALENDAR