{"count":263,"next":null,"previous":"https://memento.epfl.ch/api/v1/events/?format=json&limit=10&offset=250&ordering=lang","results":[{"id":71604,"title":"CDM-MTEI Seminars: Performance Optimization Programming: Methodology, Algorithms and Insights","slug":"cdm-mtei-seminars-performance-optimization-program","event_url":"https://memento.epfl.ch/event/cdm-mtei-seminars-performance-optimization-program","visual_url":"https://memento.epfl.ch/image/32936/200x112.jpg","visual_large_url":"https://memento.epfl.ch/image/32936/720x405.jpg","visual_maxsize_url":"https://memento.epfl.ch/image/32936/max-size.jpg","lang":"en","start_date":"2026-05-01","end_date":"2026-05-01","start_time":"11:30:00","end_time":"13:00:00","description":"<strong>Abstract</strong>:<br>\r\nThe design of efficient optimization algorithms has long relied on intuition and hand-crafted analysis. A newer paradigm -- Performance Optimization Programming (POP) -- reframes the question of algorithm analysis as a metaoptimization problem, enabling a computer-assisted, systematic search for optimal methods. We introduce Branch-and-Bound Performance Estimation Programming (BnB-PEP), a unified methodology for constructing certifiably optimal first-order methods for both convex and nonconvex optimization. Prior PEP-based approaches formulate worst-case performance as a convex semidefinite program, but this convexity is simultaneously a limitation: only a narrow class of problem setups admit such tractable reformulations. BnB-PEP overcomes this barrier by directly embracing nonconvexity. BnB-PEP is applied to several settings that were previously intractable: constructing the optimal gradient method without momentum, deriving optimal methods for reducing gradient norms of smooth nonconvex functions, and designing efficient methods for weakly convex nonsmooth optimization. In each case, the methodology yields new state-of-the-art convergence bounds. Finally, we show how the analysis generalizes to settings with gradient corruption and allows for the discovery of an entirely novel class of first-order optimization algorithms which enjoy optimal corruption protection.<br>\r\n<br>\r\n<br>\r\n<strong>References:</strong>\r\n<ul>\r\n\t<li>[Das Gupta, van Parys, Ryu: Branch-and-Bound Performance Estimation Programming](<a href=\"https://arxiv.org/pdf/2203.07305\">https://arxiv.org/pdf/2203.07305</a>)</li>\r\n\t<li>[Gösgens, van Parys: Subgradient Methods for Nonsmooth Convex Functions with Adversarial Errors](<a href=\"https://arxiv.org/abs/2510.03072\">https://arxiv.org/abs/2510.03072</a>)</li>\r\n</ul>\r\n<br>\r\n<br>\r\n<strong>Short bio: Bart van Parys</strong> is a senior researcher in the Stochastics groupt at the Dutch National Institute for Mathematics and Computer Science (CWI). Bart has played a pioneering role in developing distributionally robust optimization (DRO) as a practical tool for decision-making in uncertain environments. His doctoral work pioneered DRO formulations in stochastic control theory and power systems. Key innovations were an extension of the inequality of Gauss from 1821 in terms of tractable convex optimization problems.<br>\r\nMore recently, Bart formalized the search for statistically efficient data-driven decisions by advancing a novel meta-optimization framework via large-deviations theory. Bart's current research is primarily methodological and revolves around fundamental concerns when making decisions based on wild data sets and the automated design of better optimization algorithms.<br>\r\n ","image_description":"","creation_date":"2026-04-15T18:05:10","last_modification_date":"2026-04-15T18:05:10","link_label":"","link_url":"https://forms.cloud.microsoft/Pages/ResponsePage.aspx?id=alXC9vvEsUqix54iDfEcQ_ZC0STJ3plPhyuf4N1UEBtUQzlHRTFSUVRZOEMxTjJTU1Q5TkVQVkNLMC4uplPhyuf4N1UEBtUQzlHRTFSUVRZOEMxTjJTU1Q5TkVQVkNLMC4u&fromie=true","canceled":"False","cancel_reason":"","place_and_room":"ODY 4 03","url_place_and_room":"https://plan.epfl.ch/?room==ODY%204%2003","url_online_room":"https://epfl.zoom.us/meeting/register/-K7fsTI-QaeQE6wowwJbtw","spoken_languages":["https://memento.epfl.ch/api/v1/spoken_languages/2/?format=json"],"speaker":"Professor Bart VAN PARYS\r\n\r\nDutch National Institute for Mathematics and Computer Science (CWI)","organizer":"Prof. Daniel Kuhn","contact":"cde-seminars@epfl.ch","is_internal":"False","theme":"","vulgarization":{"id":2,"fr_label":"Public averti","en_label":"Informed public"},"registration":{"id":1,"fr_label":"Sur inscription","en_label":"Registration required"},"keywords":"","file":null,"icalendar_url":"https://memento.epfl.ch/event/export/120394/","category":{"id":1,"code":"CONF","fr_label":"Conférences - Séminaires","en_label":"Conferences - Seminars","activated":true},"academic_calendar_category":null,"domains":[],"mementos":["https://memento.epfl.ch/api/v1/mementos/6/?format=json","https://memento.epfl.ch/api/v1/mementos/27/?format=json","https://memento.epfl.ch/api/v1/mementos/38/?format=json"]},{"id":71605,"title":"NanoFrazor Workshop - Advancing Parallel Nanolithography with Decapede Technology","slug":"nanofrazor-workshop-advancing-parallel-nanolithogr","event_url":"https://memento.epfl.ch/event/nanofrazor-workshop-advancing-parallel-nanolithogr","visual_url":"https://memento.epfl.ch/image/32937/200x112.jpg","visual_large_url":"https://memento.epfl.ch/image/32937/720x405.jpg","visual_maxsize_url":"https://memento.epfl.ch/image/32937/max-size.jpg","lang":"en","start_date":"2026-04-28","end_date":"2026-04-28","start_time":"09:30:00","end_time":"16:00:00","description":"<p>Join us on April 28 for a focused workshop on advanced nanolithography with the NanoFrazor, a high-resolution thermal scanning probe lithography platform. The event will showcase the latest Decapede technology, featuring 10 parallel writing tips for significantly increased throughput without compromising nanoscale precision.<br>\r\n<br>\r\nThe morning session will include expert talks from the CTO and engineering team at  Heidelberg Instruments Nano, alongside Prof. Jürgen Brugger and researchers from the hosting laboratory. Guest speakers will present recent results and application insights enabled by the NanoFrazor. Following a joint lunch, the afternoon session will offer live tool demonstrations, in-depth technical discussions, and direct interaction with application and technology experts.</p>","image_description":"","creation_date":"2026-04-15T20:24:18","last_modification_date":"2026-04-15T20:24:18","link_label":"Registration Link","link_url":"https://forms.office.com/Pages/ResponsePage.aspx?id=l5wK4YNskEC3vlbu_MEhx01ysv1IFcRHmIkIQyJVFepUNU5ZVU1TMEdMQVNROVlRNExRNzVWVVAyRC4u&origin=QRCode","canceled":"False","cancel_reason":"","place_and_room":"BM 1130","url_place_and_room":"https://plan.epfl.ch/?room==BM%201130","url_online_room":"","spoken_languages":["https://memento.epfl.ch/api/v1/spoken_languages/2/?format=json"],"speaker":"Dr. Antoaneta Damyanova, Head of Software, Heidelberg Instruments Nano<br>\r\nProf. Dr. Jürgen Brugger, Microsystems Laboratory 1 (LMIS1), EPFL<br>\r\nEliott Clerc, R&amp;D Engineer, Heidelberg Instruments Nano\r\n\r\nMore to come","organizer":"Giulia Tagliabue\r\n\r\nJuergen Brugger\r\n\r\nElliott Clerc","contact":"giulia.tagliabue@epfl.ch\r\n\r\njuergen.brugger@epfl.ch\r\n\r\necl@himt.ch","is_internal":"False","theme":"","vulgarization":{"id":2,"fr_label":"Public averti","en_label":"Informed public"},"registration":{"id":1,"fr_label":"Sur inscription","en_label":"Registration required"},"keywords":"nanofabrication, lithography","file":null,"icalendar_url":"https://memento.epfl.ch/event/export/120395/","category":{"id":1,"code":"CONF","fr_label":"Conférences - Séminaires","en_label":"Conferences - Seminars","activated":true},"academic_calendar_category":null,"domains":[],"mementos":["https://memento.epfl.ch/api/v1/mementos/8/?format=json","https://memento.epfl.ch/api/v1/mementos/204/?format=json","https://memento.epfl.ch/api/v1/mementos/377/?format=json"]},{"id":71609,"title":"The Abstraction Trap: Lessons from a Career in the Memory Trenches","slug":"the-abstraction-trap-lessons-from-a-career-in-the","event_url":"https://memento.epfl.ch/event/the-abstraction-trap-lessons-from-a-career-in-the","visual_url":"https://memento.epfl.ch/image/32940/200x112.jpg","visual_large_url":"https://memento.epfl.ch/image/32940/720x405.jpg","visual_maxsize_url":"https://memento.epfl.ch/image/32940/max-size.jpg","lang":"en","start_date":"2026-04-29","end_date":"2026-04-29","start_time":"14:00:00","end_time":"15:00:00","description":"<div><strong><u>Abstract</u></strong><br>\r\nWhy do some foundational technologies advance rapidly while others seem frozen in time? For twenty years, production garbage collectors have remained remarkably stagnant—not for a lack of great ideas, but for a lack of the right abstractions. In this talk, I reflect on a career spent navigating the friction between theory and deployment.  I will explore how high-level abstractions—from strict encapsulation to the \"workpacket\" model—catalyze the innovation necessary to disrupt an industry status quo. Using the MMTk framework as a case study, I'll show how these architectural foundations enabled the development of systems like LXR that significantly outperform the state-of-the-art.  However, innovation is only half the battle. Drawing from recent experiences productizing MMTk for OpenJDK, Julia, and Ruby, I will reveal the hidden impediments to taking even the most promising ideas to production. I will discuss how decades of abstraction leakage in production codebases create a substantial barrier to entry for new ideas.  I'll also discuss the difficulty of motivating adoption when a lack of appropriate methodology causes the true costs of incumbent systems to be underestimated or entirely misunderstood. While this talk draws on the world of memory management, it is not a talk about garbage collection. It is a reflection on the struggle to innovate within established fields and a roadmap for using foundational abstractions to clear the way for the next generation of software.</div>\r\n\r\n<div> </div>\r\n\r\n<div><strong><u>Bio</u></strong><br>\r\nSteve Blackburn is a Senior Staff Research Scientist at Google DeepMind, a professor at the Australian National University and an ACM Fellow.  Steve has a long history in systems and programming languages with a particular interest in management and performance analysis. His current research includes hardware and software garbage collection for the fleet and mobile devices; applying AI to large scale software engineering tasks; and understanding and mitigating silent data corruption.<br>\r\n<br>\r\n<a href=\"https://www.steveblackburn.org/\">More information</a><br>\r\n </div>","image_description":"","creation_date":"2026-04-16T10:01:28","last_modification_date":"2026-04-16T10:08:37","link_label":"","link_url":"","canceled":"False","cancel_reason":"","place_and_room":"BC 410","url_place_and_room":"https://plan.epfl.ch/?room==BC%20410","url_online_room":"","spoken_languages":["https://memento.epfl.ch/api/v1/spoken_languages/2/?format=json"],"speaker":"","organizer":"","contact":"Host: Prof. Nate Foster","is_internal":"False","theme":"","vulgarization":{"id":1,"fr_label":"Tout public","en_label":"General public"},"registration":{"id":3,"fr_label":"Entrée libre","en_label":"Free"},"keywords":"","file":null,"icalendar_url":"https://memento.epfl.ch/event/export/120399/","category":{"id":1,"code":"CONF","fr_label":"Conférences - Séminaires","en_label":"Conferences - Seminars","activated":true},"academic_calendar_category":null,"domains":[],"mementos":["https://memento.epfl.ch/api/v1/mementos/6/?format=json","https://memento.epfl.ch/api/v1/mementos/27/?format=json"]}]}