Alberto Abbondandolo (Ruhr Universität Bochum)
Systolic inequalities in Hamiltonian dynamics abstract
Abstract:
I will discuss how questions and results from metric systolic geometry extend naturally to Hamiltonian dynamics, and in particular to Reeb flows. These questions and results are related to some open conjectures in symplectic geometry and in convex geometry.
10:00 • Université de Neuchâtel, Institut de Mathématiques, Rue Emile Argand 14, 218
Anna Karpova (Universität Basel)
Wasserstein metric based full waveform inversion abstract
Abstract:
Full waveform inversion (FWI) is a widely spread signal analyzing technique. For example in seismology, it is used to determine the properties of the subsurface. The inverse problem is formulated as an optimisation problem of mismatch between observed and synthetic data. Common FWI uses the least-squares (L2) norm to generate a misfit function. However, the function based on L2 norm may have several local minima. Consequently, it requires initial values that are close to the global minimum in order to avoid convergence towards a local minimum, hence opening the possibility of ending up with wrong results.
In this thesis, it has been examined whether it is an advantage to work with the quadratic Wasserstein metric instead. The Wasserstein metric corresponds to the optimal cost of rearranging one density into the other and thus recognises well the phase shift of signals. Based on a model problem, the standard FWI (with L2 norm) was compared to the FWI with quadratic Wasserstein metric. Numerical results demonstrate that the quadratic Wasserstein metric is a very promising choice of a misfit functional in seismic inversion.
12:15 • Universität Basel, Fachbereich Mathematik, Spiegelgasse 5, Room 05.002
Sylvain Lavau (Max Planck Institut, Bonn)
The Universal Lie infinity-algebroid of a singular foliation abstract
Abstract:
Any regular foliation canonically induces a \'foliation\' Lie algebroid, whose anchor map is injective. What is the similar structure for a singular foliation? We show that the natural candidate would be a Lie infinity-algebroid resolving the singular foliation. We show that such a Lie infinity-algebroid structure can be defined on any resolution of the singular foliation. This Lie infinity-algebroid is universal in thesense that any two such structures are equivalent up to homotopy. In particular this implies that their cohomologies are canonically isomorphic. Then we might have some time to discuss possible applications.
15:15 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room G 43 + Max Planck Institut, Boon
François Loeser (Université Sorbonne)
Motivic Integration: Old and New abstract
Abstract:
After providing a general introduction to motivic integration and its historical development,we will focus on some of its recent applications to geometry and representation theory.
16:15 • Université de Genève, Section Mathématiques, 2-4, rue du Lièvre, Salle 17
Felix Matthys (ITAM Businees School)
Stochastic Volatility and Model Uncertainty: Implications for Asset Prices and Optimal Portfolio Policies abstract
Abstract:
This paper presents a framework in which stochastic volatility endogenously drives model uncertainty and shows how an investor derives optimal robust consumption and portfolio policies in fully closed-form. In turbulent times, characterized by elevated volatility levels, the amount of model uncertainty the investor is facing increases and thereby rendering the estimation of his reference model more difficult. Our jump diffusive stochastic volatility model accounts for the well-known stylized facts of asset price volatility, such as volatility clustering and occasional large jumps, which has direct implications on, first, how much model uncertainty there is and second, how the investor optimally guards himself against model miss-specification. Our quantitative analysis shows that the naive investor, who has full trust in his reference model and believes volatility is constant, incurs significant wealth losses as compared to a robust investor that accounts for the time series properties of empirical volatility. We derive a semi-closed form expression for the detection-error probability which allows us to efficiently quantify the amount of model uncertainty. Moreover, we also solve the model in general equilibrium in order to determine how the equity risk premia and the risk free are affected by both, stochastic volatility and model model uncertainty. Finally, we implement a quasi-maximum likelihood estimator to fit the model to the data and find that our jump-diffusive volatility model is able to reproduce key moments of asset prices. This is joint work with Yacine Ait-Sahalia (Princeton University) and Emilio Osambela (Federal Reserve Board).
17:15 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room G 43
Anne-Sophie de Suzzoni (Université Paris 13)
[Video] Weak turbulence abstract
Abstract:
Abstract : Wave turbulence is the study of the evolution of the statistics of random waves. Weak turbulence corresponds to taking an equation, coming from hydrodynamics or quantum mechanics, which is weakly nonlinear (that is we let its nonlinearity go to zero in a certain regime). One aim of this talk is to present the first aspects of the theory of weak turbulence from a mathematical physics point of view, explain which intrinsically nonlinear behavior it describes, its link with resonances and the growth of Sobolev norms. We will then present some rigorous results and perspectives. One aim is to introduce the computational tool of Feynmann diagrams in this context, its relevance and importance in fully developing the mathematical study of weak turbulence. We will discuss a joint work with Nikolay Tzvetkov and an ongoing project with Zaher Hani.,Slides
18:10 • UZH Irchel, Winterthurerstrasse 190, Zürich, Building Y27, Room H 35/36