BEGIN:VCALENDAR VERSION:2.0 PRODID:-//Drupal iCal API//EN X-WR-CALNAME:Events items teaser X-WR-TIMEZONE:America/Toronto BEGIN:VTIMEZONE TZID:America/Toronto X-LIC-LOCATION:America/Toronto BEGIN:DAYLIGHT TZNAME:EDT TZOFFSETFROM:-0500 TZOFFSETTO:-0400 DTSTART:20250309T070000 END:DAYLIGHT BEGIN:STANDARD TZNAME:EST TZOFFSETFROM:-0400 TZOFFSETTO:-0500 DTSTART:20241103T060000 END:STANDARD END:VTIMEZONE BEGIN:VEVENT UID:686fdebdbf9f8 DTSTART;TZID=America/Toronto:20250718T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250718T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-ashwin-nayak-1 SUMMARY:Tutte colloquium-Ashwin Nayak CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Learning quantum states\n\nSPEAKER:\n Ashwin Nayak\n\nAFFILIATION:\n University of À¶Ý®ÊÓƵ\n\nLOCATION:\n MC 5501\n\nA BSTRACT: Suppose we are given a sequence of quantum registers\ninitialize d to the same quantum state rho\, and would like to learn\nthe state rho. That is\, we would like to design an algorithm that\nproduces a classica l description of an approximation to the state.\nHow many copies of rho d owe need to be able to produce a suitable\napproximation? This talk will  be a gentle introduction to the problem\nand related results.\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc05d3 DTSTART;TZID=America/Toronto:20250711T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250711T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-stephen-melczer-0 SUMMARY:Tutte colloquium-Stephen Melczer CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Automated Sequence Asymptotics\n\nSPEAKER:\n Stephen Melczer\n\nAFFILIATION:\n University of À¶Ý®ÊÓƵ\n\nLOCATION:\n MC 5501\n\nABSTRACT:Computing with any sort of object requires a way of enco ding\nit on a computer\, which poses a problem in enumerative combinatoric s\nwhere the objects of interest are (infinite) combinatorial sequences.\n Thankfully\, the generating function of a combinatorial sequence often\nsa tisfies natural algebraic/differential/functional equations\, which\ncan t hen be viewed as data structures for the sequence. In this talk\nwe survey methods to take a sequence encoded by such data structures\nand automatic ally determine asymptotic behaviour using techniques from\nthe field of an alytic combinatorics. We also discuss methods to\nautomatically characteri ze the asymptotic behaviour of multivariate\nsequences using analytic comb inatorics in several variables (ACSV).\nThe focus of each topic will be ri gorous algorithms that have already\nbeen implemented in computer algebra systems and can be easily used by\nanyone.\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc0ea0 DTSTART;TZID=America/Toronto:20250704T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250704T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-henry-wolkowicz-2 SUMMARY:Tutte colloquium-Henry Wolkowicz CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:The omega-Condition Number: Applications to P reconditioning and\nLow Rank Generalized Jacobian Updating\n\nSPEAKER:\n H enry Wolkowicz\n\nAFFILIATION:\n University of À¶Ý®ÊÓƵ\n\nLOCATION:\n MC 5501\n\nABSTRACT: Preconditioning is essential in iterative methods for\n solving linear systems. It is also the implicit objective in updating\napp roximations of Jacobians in optimization methods\, e.g.\,~in\nquasi-Newton methods. We study a nonclassic matrix condition number\,\nthe omega-condi tion number}\, omega for short. omega is the ratio of:\nthe arithmetic and geometric means of the singular values\, rather than\nthe largest and sma llest for the classical kappa-condition number. The\nsimple functions in o mega allow one to exploit  first order\noptimality conditions. We use thi s fact to derive explicit formulae\nfor (i) omega-optimal low rank updatin g of generalized Jacobians\narising in the context of nonsmooth Newton met hods\; and (ii)\nomega-optimal preconditioners of special structure for   iterative\nmethods for linear systems. In the latter context\, we analyze the\nbenefits of omega for (a) improving the clustering of eigenvalues\; ( b)\nreducing the number of iterations\; and (c) estimating the actual\ncon dition of a linear system. Moreover we show strong theoretical\nconnection s between the omega-optimal preconditioners and incomplete\nCholesky facto rizations\, and highlight the misleading effects arising\nfrom the inverse invariance of kappa. Our results confirm the efficacy\nof using the omega -condition number compared to the kappa-condition\nnumber.\n\n(Joint work with: Woosuk L. Jung\, David Torregrosa-Belen.)\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc17a9 DTSTART;TZID=America/Toronto:20250627T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250627T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-gary-au-0 SUMMARY:Tutte colloquium-Gary Au CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Worst-case instances of the stable set proble m of graphs for the\nLovász–Schrijver SDP hierarchy\n\nSPEAKER:\n Gary Au\n\nAFFILIATION:\n University of Saskatchewan\n\nLOCATION:\n MC 5501\n\n ABSTRACT:(Based on joint work with Levent Tunçel.)\n\nIn this talk\, we d iscuss semidefinite relaxations of the stable set\nproblem of graphs gener ated by the lift-and-project operator LS_+ (due\nto Lovász and Schrijver) \, and present some of our recent progress on\nthis front. In particular\, we show that for every positive integer k\,\nthe smallest graph with LS_+ -rank k contains exactly 3k vertices. This\nresult is sharp and settles a conjecture posed by Lipták and Tunçel\nfrom 2003.\n\nThe talk will be ac cessible to a general audience\, and does not assume\nany prior knowledge of lift-and-project methods.\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc1ec2 DTSTART;TZID=America/Toronto:20250620T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250620T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-sepehr-hajebi SUMMARY:Tutte colloquium-Sepehr Hajebi CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Complete bipartite induced minors (and treewi dth)\n\nSPEAKER:\n Sepehr Hajebi\n\nAFFILIATION:\n University of À¶Ý®ÊÓƵ\ n\nLOCATION:\n MC 5501\n\nABSTRACT:I will present a result that describes the unavoidable\ninduced subgraphs of graphs with a large complete biparti te induced\nminor\, and will discuss the connections and applications to b ounding\nthe treewidth in hereditary classes of graphs. If time permits\, I will\nalso sketch some proofs.\n\n Joint work with Maria Chudnovsky and Sophie Spirkl.\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc25c5 DTSTART;TZID=America/Toronto:20250613T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250613T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-rose-mccarty SUMMARY:Tutte colloquium-Rose McCarty CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:The first-order logic of graphs\n\nSPEAKER:\n  Rose McCarty\n\nAFFILIATION:\n Georgia Institute of Technology\n\nLOCAT ION:\n MC 5501\n\nABSTRACT:Over the last ten years\, many wonderful connec tions have been\nestablished between structural graph theory\, computation al complexity\,\nand finite model theory. We give an overview of this area \, focusing on\nrecent progress towards understanding the \"stable\" case. We do not\nassume any familiarity with first-order logic\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc2c7e DTSTART;TZID=America/Toronto:20250530T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250530T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-lior-gishboliner SUMMARY:Tutte colloquium-Lior Gishboliner CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Regularity lemmas for hypergraphs of bounded VC dimension:\nimproved bounds\n\nSPEAKER:\n Lior Gishboliner\, \n\nAFFILI ATION:\n University of Toronto\n\nLOCATION:\n MC 5501\n\nABSTRACT:An impor tant result at the interface of graph theory and\nlogic is that graphs of bounded VC dimension have (small) homogeneous\nvertex-partitions\, i.e.\, partitions where almost every pair of parts\nhas density close to 0 or 1. Recently\, Chernikov and Towsner proved a\nhypergraph generalization of th is fact. The quantitative aspects of\ntheir result remain open. I will pre sent some recent progress on this\nproblem\, answering two questions of Te rry. This is a joint work with\nAsaf Shapira and Yuval Wigderson.\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc3311 DTSTART;TZID=America/Toronto:20250523T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250523T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-david-torregrossa-belen SUMMARY:Tutte colloquium-David Torregrossa Belén CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Splitting algorithms for monotone inclusions with\nminimal dimension\n\nSPEAKER:\n David Torregrossa Belén\n\nAFFILIA TION:\n Center for Mathematical Modeling\, University of Chile\n\nLOCATIO N:\n MC 5501\n\nABSTRACT: Many situations in convex optimization can be mo deled as the\nproblem of finding a zero of a monotone operator\, which ca n be\nregarded as a generalization of the gradient of a differentiable\nc onvex function. In order to numerically address this monotone\ninclusion problem\, it is vital to be able to exploit the inherent\nstructure of th e operator defining it. The algorithms in the family\nof the splitting me thods achieve this by iteratively solving simpler\nsubtasks that are defi ned by separately using some parts of the\noriginal problem.\n\nIn the fi rst part of this talk\, we will introduce some of the\nmost relevant mono tone inclusion problems and present their\napplications to optimization. Subsequently\, we will draw our\nattention to a common anomaly that has p ersisted in the design of\nmethods in this family: the dimension of the u nderlying space\n—which we denote as lifting— of the algorithms abnor mally\nincreases as the problem size grows. This has direct implications on\nthe computational performance of the method as a result of the\nincre ase of memory requirements. In this framework\, we characterize\nthe mini mal lifting that can be obtained by splitting algorithms\nadept at solvin g certain general monotone inclusions. Moreover\, we\npresent splitting m ethods matching these lifting bounds\, and\nthus having minimal lifting.\ n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc3eb7 DTSTART;TZID=America/Toronto:20250516T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250516T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-michael-borinsky SUMMARY:Tutte colloquium-Michael Borinsky CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Constraining moduli space cohomology by count ing graphs\n\nSPEAKER:\n Michael Borinsky\n\nAFFILIATION:\n Perimeter Inst itute\n\nLOCATION:\n MC 5501\n\nABSTRACT: In 1992\, Kontsevich defined com plexes spanned by graphs.\nThese \ncomplexes are increasingly prominent i n algebraic topology\,\ngeometric \ngroup theory and mathematical physics . For instance\, a 2021 theorem\nby \nChan-Galatius and Payne implies tha t the top-weight cohomology of\nthe \nmoduli space of curves of genus g i s equal to the homology of a\nspecific \ngraph complex. I will present a new theorem on the asymptotic growth \nrate of the Euler characteristic o f this graph complex and explain\nits \nimplication on the cohomology of the moduli space of curves. The\nproof \ninvolves solving a specific grap h counting problem.\n\n \n DTSTAMP:20250710T153941Z END:VEVENT BEGIN:VEVENT UID:686fdebdc4a32 DTSTART;TZID=America/Toronto:20250509T153000 SEQUENCE:0 TRANSP:TRANSPARENT DTEND;TZID=America/Toronto:20250509T163000 URL:/combinatorics-and-optimization/events/tutte-colloq uium-luke-schaeffer SUMMARY:Tutte colloquium-Luke Schaeffer CLASS:PUBLIC DESCRIPTION:Summary \n\nTITLE:Faster linear algebra using treewidth\n\nSPEA KER:\n Luke Schaeffer\n\nAFFILIATION:\n University of À¶Ý®ÊÓƵ\n\nLOCATION :\n MC 5501\n\nABSTRACT: \n\n: We look at the complexity of solving sparse linear systems as a\nfunction of the treewidth of the instance. That is\, the sparse matrix\nis associated with a sparse graph\, and solutions can be found faster\nwhen that graph has low treewidth. We give a parameterize d algorithm\nin system size and treewidth achieving the conjectured optima l\nperformance.\n\nThis is joint work with Daniel Grier.\n\n \n DTSTAMP:20250710T153941Z END:VEVENT END:VCALENDAR