BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CMSA - ECPv6.15.20//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:CMSA
X-ORIGINAL-URL:https://cmsa.fas.harvard.edu
X-WR-CALDESC:Events for CMSA
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20250309T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20251102T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20260308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20261101T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20270314T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20271107T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260518T090000
DTEND;TZID=America/New_York:20260522T170000
DTSTAMP:20260508T211317
CREATED:20250623T220157Z
LAST-MODIFIED:20260508T164511Z
UID:10003754-1779094800-1779469200@cmsa.fas.harvard.edu
SUMMARY:Workshop on Calabi-Yau metrics and optimal transport
DESCRIPTION:Workshop on Calabi-Yau metrics and optimal transport \nDates: May 18–22\, 2026 \nLocation: Harvard CMSA\, 20 Garden Street\, Cambridge MA \nRecent advances in the study of Calabi-Yau metrics have revealed an interesting connection with optimal transport\, and the regularity theory for optimal transport is expected to play an increasingly important role in the study of Kähler geometry. The goal of this workshop is to bring together the optimal transport and complex geometry communities to investigate problems arising from these exciting developments. \nRegister to attend in-person \nRegister for Zoom Webinar \n  \nMinicourse Speakers \n\nRobert McCann\, University of Toronto\nYang Li\, Cambridge University\n\nWorkshop Speakers \n\nRolf Andreasson\, Chalmers University\, Sweden\nBenjy Firester\, MIT\nJakob Hultgren\, Umea University\, Sweden\nYoung-Heon Kim\, University of British Columbia\nNam Le\, Indiana University\nJiakun Liu\, University of Sydney\nArghya Rakshit\, University of Toronto\nGabor Szekelyhidi\, Northwestern University\nYueqiao Wu\, Johns Hopkins University\n\nOrganizers: \n\nTristan Collins\, University of Toronto\nMattias Jonsson\, University of Michigan\nConnor Mooney\, University of California\, Irvine\nFreid Tong\, University of Toronto\n\n  \nSchedule (subject to change) \nMonday\, May 18\, 2026 \n9:00–9:30 am\nBreakfast \n9:30–10:45 am\nTutorial: Yang Li\, Cambridge University (via Zoom Webinar) \n10:45–11:15 am\nBreak \n11:15 am–12:30 pm\nTutorial: Robert McCann\, University of Toronto\nTitle: A geometric approach to apriori estimates for optimal transport maps\nAbstract: A key inequality which underpins the regularity theory of optimal transport for costs satisfying the Ma-Trudinger-Wang condition is the Pogorelov second derivative bound. This translates to an a priori interior modulus of the differential estimate for smooth optimal maps. We describe a new derivation of this estimate with Brendle\, Leger and Rankin which relies in part on Kim\, McCann\, and Warren’s observation that the graph of an optimal map becomes a volume maximizing non-timelike submanifold when the product of the source and target domains is endowed with a suitable pseudo-Riemannian geometry that combines both the marginal densities and the cost. This unexpected links optimal transport to the plateau problem in geometry with split signature\, and shows the key difficulty is showing the maximizing non-timelike submanifold is in fact (uniformly) spacelike. J. Reine Angew. Math. 817 (2024) 251-266 doi.org/10.1515/crelle-2024-0071 arXiv 2311.10208 \n12:30–2:00 pm\nLunch (catered) \n2:00–3:15 pm\nTalk: Nam Le\, Indiana University\nTitle: Variational approach to degenerate Monge-Ampère equations with mixed measures and monotonicity\nAbstract: In this talk\, we will discuss the solvability and uniqueness for several degenerate Monge-Ampère equations including the Monge-Ampère eigenvalue problem in real Euclidean spaces that involve singular Borel measures. Our approach systematically analyzes the Monge-Ampère energy from the variational point of view and appropriately exploits monotonicity arguments. We will examine several essential tools: the mixed Monge-Ampère measure\, Aleksandrov-Blocki-Jerison type maximum principles\, convex envelope\, comparison principles for subcritical equations\, and integration by parts whose failure leads to symmetry breaking and nonuniqueness phenomena. \n3:15–3:45 pm\nBreak \n3:45–5:00 pm\nTalk: Yueqiao Wu\, Johns Hopkins University\nTitle: Valuative aspects of complete Calabi-Yau metrics of Euclidean volume growth\nAbstract: The search of a complete Calabi-Yau metric on an affine variety X amounts to solving a complex Monge-Ampère equation subject to nice “boundary conditions” at infinity. In the case where X is the complement of an SNC anticanonical divisor on a Fano manifold\, generalizing the work of Tian-Yau\, Collins-Li showed that such boundary data can be extracted from solutions to certain real Monge-Ampère equations. If we require the metric to have Euclidean volume growth\, however\, it is understood that the boundary conditions should come from prescribing a Calabi-Yau asymptotic cone at infinity. This is the same as giving the algebro-geometric data of a valuation which induces a degeneration of X to a K-stable affine cone. In this talk\, we will explain that such valuations in fact always come from Fano type compactifications of X\, similar to the ones considered by Tian-Yau and Collins-Li. In addition\, K-semistability of the affine cone can be characterized intrinsically by a valuative criterion on X. Based on joint work with Mattias Jonsson. \n  \nTuesday\, May 19\, 2026 \n9:00–9:30 am\nBreakfast \n9:30–10:45 am\nTutorial: Robert McCann\, University of Toronto\nTitle: Trading linearity for ellipticity: A low regularity Lorentzian splitting theorem\nAbstract: While Einstein’s theory of gravity is formulated in a smooth setting\, the celebrated singularity theorems of Hawking and Penrose describe many physical situations in which this smoothness must eventually breakdown. It is thus of great interest to study the theory in low regularity settings. In the lecture\, we establish a low regularity splitting theorem by sacrificing linearity of the d’Alembertian to recover ellipticity. We exploit a negative homogeneity $p$-d’Alembert operator for this purpose. The same technique yields a simplified proof of Eschenberg (1988) Galloway (1989) and Newman’s (1990) confirmation of Yau’s (1982) conjecture\, bringing all three Lorentzian splitting results into a framework closer to the Cheeger-Gromoll splitting theorem from Riemannian geometry. Based on joint work with Mathias Braun\, Nicola Gigli\, Argam Ohanyan\, and Clemens Saemann: 1) arXiv 2501.00702 2) arXiv 2408.15968 3) arXiv 2410.12632 4) arXiv 2507.06836 \n10:45–11:15 am\nBreak \n11:15 am–12:30 pm\nTutorial: Yang Li\, Cambridge University (via Zoom Webinar) \n12:30–2:00 pm\nLunch Break \n2:00–3:15 pm\nTalk: Young-Heon Kim\, University of British Columbia \n3:15–3:45 pm\nBreak \n3:45–5:00 pm\ntba \n6:30 pm\nDinner \n  \nWednesday\, May 20\, 2026 \n9:00–9:30 am\nBreakfast \n9:30–10:45 am\nTutorial: Yang Li\, Cambridge University (via Zoom Webinar) \n10:45–11:15 am\nBreak \n11:15 am–12:30 pm\nTutorial: Robert McCann\, University of Toronto\nTitle: The monopolist’s free boundary problem in the plane: an excursion into the economic value of private information\nAbstract: The principal-agent problem is an important paradigm in economic theory for studying the value of private information: the nonlinear pricing problem faced by a monopolist is one example; others include optimal taxation and auction design. For multidimensional spaces of consumers (i.e. agents) and products\, Rochet and Chone (1998) reformulated this problem as a concave maximization over the set of convex functions\, by assuming agent preferences are bilinear in the product and agent parameters. This optimization corresponds mathematically to a convexity-constrained obstacle problem. The solution is divided into multiple regions\, according to the rank of the Hessian of the optimizer.\nIf the monopolists costs grow quadratically with the product type we show that a partially smooth free boundary delineates the region where it becomes efficient to customize products for individual buyers. We give the first complete solution of the problem on square domains\, and discover new transitions from unbunched to targeted and from targeted to blunt bunching as market conditions become more and more favorable to the seller.\nBased on works with Kelvin Shuangjian Zhang\, Cale Rankin\, and Lucas O’Brien in various combinations:\n1) Math. Models Methods Appl. Sci. 34 (2024) 2351-2394; 2) J. Convex Anal. (Rockafellar 90 Issue)\, 32 (2) (2025) 579-584; 3) arXiv 2303.04937; 4) arxiv 2412.15505; 5) arXiv 2603.14100. \n  \nThursday\, May 21\, 2026 \n9:00–9:30 am\nBreakfast \n9:30–10:45 am\nTalk: Gabor Szekelyhidi\, Northwestern University \n10:45–11:15 am\nBreak \n11:15 am–12:30 pm\nTalk: Rolf Andreasson\, Chalmers University\, Sweden\nTitle: Optimal transport between boundaries of dual reflexive polytope\nAbstract: I will present an optimal transport problem between the boundaries of a pair of reflexive polytopes. Under a certain structural condition on its solution\, this problem is related the study of metric degenerations of families of Calabi–Yau hypersurfaces in the corresponding toric Fano variety. A better understanding of such solutions and their regularity would shed light on several aspects of the degeneration and conjectural Gromov–Hausdorff limit\, and I will present some open directions of research. This is based on joint work with Jakob Hultgren\, Mattias Jonsson\, Enrica Mazzon and Nicholas McCleerey. \n12:30–2:00 pm\nLunch Break \n2:00–3:15 pm\nTalk: Jakob Hultgren\, Umea University\, Sweden \n3:15–3:45 pm\nBreak \n3:45–5:00 pm\nTalk: Benjy Firester\, MIT \n  \nFriday\, May 22\, 2026 \n9:00–9:30 am\nBreakfast \n9:30–10:45 am\nTalk: Jiakun Liu\, University of Sydney\nTitle: Free boundary problems in optimal transportation\nAbstract: In this talk\, I will present some recent results on the regularity of free boundaries in optimal transportation\, including higher-order regularity\, global regularity\, and a model case involving multiple targets. These results are based on a series of joint works with Shibing Chen\, Xianduo Wang\, and Xu-Jia Wang. \n10:45–11:15 am\nBreak \n11:15 am–12:30 pm\nTalk: Arghya Rakshit\, University of Toronto\nTitle: Solutions to the Monge–Ampère equation with singular structures\nAbstract: We construct examples of solutions to the Monge–Ampère equation with point masses exhibiting polyhedral singular structures. We further analyze the stability of these singular sets under small perturbations of the data. In addition\, we construct solutions whose Monge–Ampère measure contains a singular component supported on lower-dimensional sets and we study the regularity of such solutions. \n 
URL:https://cmsa.fas.harvard.edu/event/cymetrics/
LOCATION:CMSA 20 Garden Street Cambridge\, Massachusetts 02138 United States
CATEGORIES:Workshop
ATTACH;FMTTYPE=image/png:https://cmsa.fas.harvard.edu/media/CY-Workshop.2.png
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260520T140000
DTEND;TZID=America/New_York:20260520T150000
DTSTAMP:20260508T211317
CREATED:20260429T133019Z
LAST-MODIFIED:20260429T143145Z
UID:10003942-1779285600-1779289200@cmsa.fas.harvard.edu
SUMMARY:Separation of timescales controls feature learning and overfitting in large neural networks
DESCRIPTION:New Technologies in Mathematics Seminar \nSpeaker: Pierfrancesco Urbani\, Universite Paris-Saclay\, CNRS\, CEA\, Institut de physique theorique \nTitle: Separation of timescales controls feature learning and overfitting in large neural networks \nAbstract: To understand the inductive bias and generalization capabilities of large\, overparameterized machine learning models\, it is essential to analyze the dynamics of their training algorithms. Using dynamical mean field theory we investigate the learning dynamics of large two-layer neural networks. Our findings reveal that\, for networks with a large width\, the training process exhibits a separation of timescales phenomenon. This leads to several key observations:\n1. The emergence of a slow timescale linked to the growth in Gaussian/Rademacher complexity of the network;\n2. An inductive bias favoring low complexity when the initial model complexity is sufficiently small;\n3. A dynamical decoupling between feature learning and overfitting phases;\n4. A non-monotonic trend in test error\, characterized by a “feature unlearning” regime at later stages of training.\nJoint work with Andrea Montanari. \n  \n 
URL:https://cmsa.fas.harvard.edu/event/newtech_52026/
LOCATION:Virtual
CATEGORIES:New Technologies in Mathematics Seminar
ATTACH;FMTTYPE=image/png:https://cmsa.fas.harvard.edu/media/CMSA-NTM-Seminar-5.20.2026.docx.png
END:VEVENT
END:VCALENDAR