Status  Confirmed 
Seminar Series  SOUTENTH 
Subjects  physics 
Date  Thursday 18 October 2018 
Time  14:00 
Institute  IPHT 
Seminar Room  Amphi Claude Bloch, Bât. 774 
Speaker's Last Name  Christian Schmidt 
Speaker's First Name  
Speaker's Email Address  
Speaker's Institution  IPhT 
Title  Statistical Physics of Sparse and Dense Models in Optimization and Inference 
Abstract  Datasets come in a variety of forms and from a broad range of different applications. Typically, the observed data is noisy or in some other way subject to randomness. The recent developments in machine learning have revived the need for exact theoretical limits of probabilistic methods that recover information from noisy data. \par In this thesis we are concerned with the following two questions. \\ What is the asymptotically best achievable performance? \\ And how can this performance be achieved, i.e., what is the optimal algorithmic strategy? \par The above questions can be studied in a probabilistic framework, which leads to an average (i.e. typical) case answer. Such a probabilistic formulation is natural to statistical physics and leads to a formal analogy with problems in disordered systems. In turn, this permits to harvest the methods developed in the study of disordered systems, to attack constraint satisfaction and statistical inference problems. \par I will present four contributions. First, a statistical physics investigation of the circular coloring problem is carried out that reveals several distinct features. Second, new rigorous upper bounds on the size of minimal contagious sets in random graphs, with bounded maximum degree, are obtained. Third, the phase diagram of the dense DawidSkene model is derived by mapping the problem onto lowrank matrix factorization. The associated approximate message passing algorithm is evaluated on synthetic and realworld data. Finally, we propose an approach to derive the Bayes optimal denoising mean square error for a restricted class of extensive rank matrix estimation problems. 
arXiv Preprint Number  
Comments  
Attachments 
To Generate a poster for this seminar : [ Postscript  PDF ]

[ English version ] 