Geometric Analysis Seminar -- Some new solutions to the Strominger system

11/25/2015 - 13:30
11/25/2015 - 14:30
Speaker: 
Teng Fei (MIT)
Location: 
Burnside Hall, Rm. 920 (McGill University)
Abstract: 

The Strominger system is a system of PDEs derived by Strominger in his study of compactification of heterotic strings with torsion. It can be thought of as a generalization of Ricci-flat metrics on non-Kahler Calabi-Yau 3-folds. We present some new solutions to the Strominger system on a class of noncompact Calabi-Yau 3-folds constructed by twistor technique. These manifolds include the resolved conifold Tot(O(-1,-1)->P^1) as a special case.

 

 

Last edited by on Wed, 11/25/2015 - 09:59

Geometry-Topology -- Octonionic-algebraic structure and curvature of the moduli space of $G_2$ manifolds

11/27/2015 - 11:00
11/27/2015 - 12:00
Speaker: 
Spiro Karigiannis (University of Waterloo)
Location: 
PK-5115, Pavillon Président-Kennedy, UQAM
Last edited by on Thu, 11/26/2015 - 10:08

Seminar Probabilités -- Moments, rough initial data and intermittency for SPDEs

11/26/2015 - 16:30
11/26/2015 - 17:30
Speaker: 
Le Chen (Kansas University)
Location: 
Burnside Hall, Rm. 920 (McGill University)
Abstract: 

Intermittency refers to the property that solution exhibits both tall peaks and low valleys. Zeldovich et al. observed that Intermittency is a very universal phenomenon which occurs practically irrespective of detailed properties of the background instability in a random medium provided only that the random field is of multiplicative type … Our goal is to turn this qualitative statement into quantitative theorems for various stochastic partial differential equations (SPDEs) subject to some noise of multiplicative type. In particular, I will first present, in this talk, some results on the stochastic heat equation (SHE). A key tool in this study is a sharp moment formula, which enables us to establish existence of a random field solution starting from measure-valued initial data. Moreover, when initial data is localized, using this moment formula, we are able to establish some exact propagation speeds of these tall peaks, the so-called intermittency fronts or growth indices, certain quantities defined and first studied by Conus and Khoshnevisan. Then I will show that techniques used to study SHE can be applied to various other SPDEs, such as the stochastic wave equation, SHE with the Laplacian replaced by a fractional Laplacian, and SHE on R^d with a spatially colored noise. This talk is based on some joint works with Robert C. Dalang and Kunwoo.

Last edited by on Mon, 11/23/2015 - 17:32

Working Seminar Mathematical Physics -- The physics of the Seiberg-Witten curve

11/26/2015 - 16:00
11/26/2015 - 17:00
Speaker: 
Patrick Labelle (Bishop’s, Sherbrooke)
Location: 
Concordia University, Math Help Center, room 912.00 / Library Building, 1400 de Maisonneuve Blvd. West, Montréal
Abstract: 

In 1994, Seiberg and Witten used a clever approach  to obtain the low energy dynamics of a certain supersymmetric gauge theory.  Using monodromy data obtained from physical considerations they observed  that the physics was encoded in an elliptic curve. The talk will describe the  physics problem and the solution.  It will also briefly address the connection  with the partition function of Nekrasov and a conjectured relation to the  topological recursion of Eynard and Orantin.

Last edited by on Mon, 11/23/2015 - 17:30

Colloquium DIRO -- Big Data and Integer Programming

11/26/2015 - 15:30
11/26/2015 - 16:30
Speaker: 
Andrea Lodi (École Polytechnique)
Location: 
Université de Montréal, Pavillon André-Aisenstadt, 2920 ch. de la Tour, salle 3195
Abstract: 

In this talk I review a couple of applications on Big Data that I personally like and I try to explain my point of view as a Mathematical Optimizer---especially concerned with discrete (integer) decisions---on the subject. I advocate a tight integration of Data Mining, Machine Learning and Mathematical Optimization (among others) to deal with the challenges of decision-making in Data Science. Those challenges are the core of the mission of the Canada Excellence Research Chair in "Data Science for Real-time Decision Making” that I hold.

Last edited by on Mon, 11/23/2015 - 17:29

Inference regarding within-family association in disease onset times under biased sampling schemes

11/26/2015 - 15:30
11/26/2015 - 16:30
Speaker: 
Richard Cook (University of Waterloo)
Location: 
Burnside Hall, Rm. 306 (McGill University)
Abstract: 
In preliminary studies of the genetic basis for chronic conditions, interest routinely  lies in the within-family dependence in disease status. When probands are selected from disease registries and their respective families are recruited, a variety of ascertainment   bias-corrected methods of inference are available which are typically based on models for correlated binary data. This approach ignores the age that family members are at the time of assessment. We consider copula-based models for assessing the within-family dependence   in the disease onset time and disease progression, based on right-censored and current status observation of the non-probands. Inferences based on likelihood, composite likelihood and estimating functions are each discussed and compared in terms of asymptotic and empirical  relative efficiency. This is joint work with Yujie Zhong.
 
Last edited by on Thu, 11/26/2015 - 10:07

Seminar Geometric Group Theory -- Metrics on the knot concordance space

11/25/2015 - 15:00
11/25/2015 - 16:00
Speaker: 
Mark Powell (UQAM)
Location: 
Burnside Hall, Rm. 920 (McGill University)
Abstract: 

The slice genus of a knot is the minimal genus of a smoothly embedded surface in the 4-ball whose boundary is the knot. This can be used to define an integer valued metric on the space of knots up to concordance. I will describe a refinement of this that uses a 2-complex construction called a grope (roughly, a grope is tower of embedded surfaces) to better approximate a slice disc, leading to a rational valued metric that refines the slice genus metric and reveals non-discrete behaviour.

Last edited by on Mon, 11/23/2015 - 17:26

Lecture Club mathématique UdeM -- Voyage au monde du dilogarithme

11/25/2015 - 12:30
11/25/2015 - 13:30
Speaker: 
Matilde Lalín (Université de Montréal)
Location: 
Université de Montréal, Pavillon Claire-McNicoll, salle Z-255
Abstract: 

Dans cette présentation, nous allons parler du dilogarithme, une fonction curieuse qui est reliée a la fonction zêta de Riemann, et qui se présente dans plusieurs endroits de la théorie des nombres et même de la topologie.

Last edited by on Mon, 11/23/2015 - 17:22

Working Seminar Mathematical Physics -- Statistical properties of some fast-slow dynamical systems

11/24/2015 - 16:00
11/24/2015 - 19:00
Speaker: 
Jacopo de Simoi (University of Toronto)
Location: 
Burnside Hall, Rm. 920 (McGill University)
Last edited by on Mon, 11/23/2015 - 17:19

Seminar Statistique Sherbrooke -- Tests statistiques pour la détection de changements graduels dans les séries chronologiques

11/24/2015 - 15:30
11/24/2015 - 16:30
Speaker: 
Jean-François Quessy (UQTR)
Location: 
Salle D4-2019, Université de Sherbrooke 2500, boul. de l'Université, Sherbrooke (Québec)
Abstract: 

Pouvoir identifier formellement la présence de changements de régime dans une série est un aspect important de l’analyse de données chronologiques, notamment dans des domaines tels la finance, l’hydrologie et la climatologie. Généralement, les procédures statistiques disponibles supposent qu’un changement éventuel se produit de façon abrupte au sens où le comportement stochastique se modifie «radicalement» en un temps K indéterminé. Cette hypothèse d’un changement abrupt est toutefois irréaliste dans plusieurs applications, en particulier dans les données environnementales, où les changements observés ont plutôt un caractère graduel; autrement dit, les changements se produisent doucement entre des temps K1 et K2. Dans cet exposé, je proposerai d’abord un modèle à changements graduels qui généralise le cas abrupt. Je montrerai ensuite comment construire des statistiques de test convergentes et des estimateurs des temps de changement à partir de ce modèle et étudierai leur comportement asymptotique. Je montrerai aussi comment calculer des valeurs critiques valides pour les tests en utilisant une version sérielle de la méthode du multiplicateur. Je présenterai ensuite quelques résultats d’une étude de simulation afin d’étudier le comportement à tailles finies des tests et des estimateurs sous divers scénarios de séries chronologiques à une et deux dimensions. Je terminerai mon exposé par une analyse de données climatologiques et discuterai de possibles extensions de mon approche.

Last edited by on Mon, 11/23/2015 - 17:16