BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Πανεπιστήμιο Κρήτης - ECPv6.16.2//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-ORIGINAL-URL:https://www.uoc.gr/
X-WR-CALDESC:Events for Πανεπιστήμιο Κρήτης
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:Europe/Athens
BEGIN:DAYLIGHT
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:EEST
DTSTART:20240331T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:EET
DTSTART:20241027T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:EEST
DTSTART:20250330T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:EET
DTSTART:20251026T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:EEST
DTSTART:20260329T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:EET
DTSTART:20261025T010000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Europe/Athens:20250908T120000
DTEND;TZID=Europe/Athens:20250908T140000
DTSTAMP:20260610T213610
CREATED:20250829T063732Z
LAST-MODIFIED:20250829T063732Z
UID:10001617-1757332800-1757340000@www.uoc.gr
SUMMARY:Ομιλία με θέμα: Minimizing Information for Inference: Structural and Noisy Query Models\, Charalampos Tsourakakis\, RelationalAI and Boston University\, 8/9/2025\, 12:00\, K206
DESCRIPTION:Title: Minimizing Information for Inference: Structural and Noisy Query Models \nSpeaker: Charalampos Tsourakakis\, RelationalAI and Boston University \nLocation: “K206\, Computer Science Department\, University of Crete”\nDate: Monday\, September 8\, 2025\nTime: 12:00 \nzoom: https://uoc-gr.zoom.us/j/85149487613 \nHost: Yannis Tollis \nAbstract: \nWe study two complementary paradigms of inference under limited information: one governed by deterministic structural constraints\, the other by noisy\, probabilistic feedback. In the first setting\, we introduce the Targeted Least Cardinality Candidate Key (TCAND) problem\, which seeks the smallest set of variables determining a target set under a given system of functional dependencies. This problem generalizes the long-standing classical problem of candidate key discovery. We shed new light on its computational complexity and approximability\, highlighting structural challenges that arise even in natural generalizations. \nIn contrast\, the second problem addresses joint alignment under a faulty oracle: given noisy observations of pairwise differences between discrete variables\, the goal is to recover the global alignment using as few queries as possible. We design a time-optimal\, non-adaptive algorithm that provably matches lower bounds on query complexity and significantly simplifies and improves upon a recent approach by Chen and Candès. Our method offers both theoretical optimality and practical simplicity\, advancing the state of the art in alignment under noise. \nTogether\, these works examine the interplay between structure\, noise\, and limited information access in inference problems—whether in minimizing attribute sets implied by functional dependencies or in learning latent variables through constrained\, noisy observations. \nThis is joint work with Kasper Green Larsen\, Michael Mitzenmacher\, Vasileios Nakos and Hung Ngo. \nBio: Dr. Charalampos Tsourakakis earned his Ph.D. from the Algorithms\, Combinatorics\, and Optimization (ACO) program at Carnegie Mellon University\, and subsequently held postdoctoral positions at Harvard and Brown Universities. He holds a Diploma in Electrical and Diploma Engineering from the National  Technical University of Athens and a Master of Science from the Machine  Learning Department at Carnegie Mellon University. Before joining Boston University\, he worked as a researcher in the Google Brain team. He won a best paper award in IEEE Data Mining\, has delivered three tutorials in  the ACM SIGKDD Conference on Knowledge Discovery and Data Mining\, and has designed two graph mining libraries for large-scale graph mining\, one of which has been officially included in Windows Azure. Currently\, he is building a state-of-the-art knowledge graph system at RelationalAI\, where he contributes to core innovations in query optimization\, relational inference\, and graph-native computation for enterprise-scale applications. His research focuses on the design of scalable algorithms and machine learning methods for analyzing large-scale datasets\, with emphasis on network data\, and applications across data science\, finance\, and computational social science.
URL:https://www.uoc.gr/event/%ce%bf%ce%bc%ce%b9%ce%bb%ce%af%ce%b1-%ce%bc%ce%b5-%ce%b8%ce%ad%ce%bc%ce%b1-minimizing-information-for-inference-structural-and-noisy-query-models-charalampos-tsourakakis-relationalai-and-boston-un/
CATEGORIES:Διάλεξη
END:VEVENT
END:VCALENDAR