{"id":652,"date":"2019-03-01T08:49:36","date_gmt":"2019-03-01T08:49:36","guid":{"rendered":"http:\/\/communication-eng.unipr.it\/?p=652"},"modified":"2021-04-09T09:24:35","modified_gmt":"2021-04-09T09:24:35","slug":"652","status":"publish","type":"post","link":"http:\/\/communication-eng.unipr.it\/index.php\/2019\/03\/01\/652\/","title":{"rendered":"Ph. D course on &#8220;Optimization using Graphical Models&#8221;"},"content":{"rendered":"<p style=\"text-align: left;\" align=\"center\">We announce a course for Ph. D students on &#8220;Optimization using Graphical Models&#8221;, March 5-14, 2019.<\/p>\n<h1 class=\"western\" align=\"center\"><span style=\"font-size: medium;\"><span lang=\"en-US\"><u><b>Ph.D. Course in \u201cOptimization using Graphical Models\u201d<\/b><\/u><\/span><\/span><\/h1>\n<h1 class=\"western\" align=\"center\"><span style=\"font-size: medium;\"><span lang=\"en-US\"><u><b>S.D.I.A. &#8211; Universit\u00e0 di Parma<\/b><\/u><\/span><\/span><\/h1>\n<h5 style=\"text-align: center;\">March 5 &#8211; 14, 2019<\/h5>\n<h1 class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\"><b>Syllabus<\/b><\/span><\/span><\/h1>\n<h3 class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\"><i>Topic 1: Basics on convex optimization (Locatelli)<\/i><\/span><\/span><\/h3>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\">Definitions of convex functions and sets. Brief sketches about complexity and algorithms. Constraint qualifications. Optimality conditions (KKT). Lagrangian duality.<\/span><\/span><\/p>\n<h3 class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\"><i>Topic 2: Bayesian inference and graphical models <\/i><\/span><\/span><span style=\"font-size: medium;\"><span lang=\"it-IT\"><i>(Colavolpe and Vannucci)<\/i><\/span><\/span><\/h3>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\">Bayesian Networks and Markov Random Fields: Inference in general Graphs.<\/span><\/span><\/p>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\">Variational Inference techniques in Machine Learning and Artificial Intelligence: Belief Propagation (BP) and Loopy Belief Propagation<\/span><\/span><\/p>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"it-IT\">Factor graphs and sum-product algorithm: general framework and applications to communications.<\/span><\/span><\/p>\n<h3 class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"it-IT\"><i>Topic 3: Variational Inference and the Free Energies methods of Physics (Vannucci)<\/i><\/span><\/span><\/h3>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"it-IT\">Approximate Inference and factorized distributions. Variational mixtures of Gaussians. Exponential Family distributions. Variational message passing. Expectation Propagation and its implementation on graphs. <\/span><\/span><\/p>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"it-IT\">Energy functions and their minimization schemes. Variational average energy and variational entropy; Gibbs and Helmholtz free energies. Stationary conditions for the Bethe free energy and its connections with Loopy Belief Propagation. The Mean Field approximation.<\/span><\/span><\/p>\n<h3 class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\"><i>Topic 4: Graphical models for optimization &#8211; Trajectory planning and power control (Consolini) <\/i><\/span><\/span><\/h3>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\">Speed and trajectory planning problems, dynamic programming. Power control in radio systems. Reduction of previous problems to a standard form and definition of the associated graph. Graph-based solution algorithms.<\/span><\/span><\/p>\n<h3 class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\"><i>Topic 5: The Physics of Statistical Inference Problems (Caire) <\/i><\/span><\/span><\/h3>\n<p class=\"western\" align=\"justify\"><span style=\"font-size: medium;\"><span lang=\"en-US\">Analysis of inference problems (in particular, posterior-mean estimation and MAP detection) using the Replica method of statistical physics. Decoupled channel model interpretation.<\/span><\/span><\/p>\n<h3><span style=\"font-size: medium;\"><span lang=\"en-US\"><i>Topic 6: Sparse Signal Reconstruction, Aka Compressed Sensing (Caire)<\/i><\/span><\/span><\/h3>\n<p><span style=\"font-size: medium;\"><span lang=\"en-US\"><br \/>\nProblem formulation. Convex relaxation, L1 and LASSO. Performance guarantees. The AMP algorithm and its state evolution analysis. Variations on the theme: multiple measurement vector (MMV) and Vector-AMP (VAMP), Non-Negative Least-Squares, and a new Maximum Likelihood approach. Some applications:<\/span><\/span><\/p>\n<p><span style=\"font-size: medium;\"><span lang=\"en-US\"><br \/>\ni) channel estimation and pilot decontamination in massive MIMO with correlated antennas; <\/span><\/span><\/p>\n<p><span style=\"font-size: medium;\"><span lang=\"en-US\">ii) SPARC codes and their decoding algorithms;<\/span><\/span><\/p>\n<p><span style=\"font-size: medium;\"><span lang=\"en-US\"> iii) Activity detection and large-scale path loss estimation in massive random access.<\/span><\/span><\/p>\n<h3><span style=\"font-size: medium;\"><span lang=\"en-US\"><i>Topic 7: Some Relations Between Replica-analysis and Amp State Evolution<\/i><\/span><\/span><\/h3>\n<p><span style=\"font-size: medium;\"><span lang=\"en-US\"> (recovering and rigorizing the replica method subject to certain model assumptions).<\/span><\/span><\/p>\n<p class=\"western\" align=\"justify\">\n","protected":false},"excerpt":{"rendered":"<p>We announce a course for Ph. D students on &#8220;Optimization using Graphical Models&#8221;, March 5-14, 2019. Ph.D. Course in \u201cOptimization using Graphical Models\u201d S.D.I.A. &#8211; Universit\u00e0 di Parma March 5 &#8211; 14, 2019 Syllabus Topic 1: Basics on convex optimization (Locatelli) Definitions of convex functions and sets. Brief sketches&hellip;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-652","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/posts\/652","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/comments?post=652"}],"version-history":[{"count":7,"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/posts\/652\/revisions"}],"predecessor-version":[{"id":659,"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/posts\/652\/revisions\/659"}],"wp:attachment":[{"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/media?parent=652"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/categories?post=652"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/communication-eng.unipr.it\/index.php\/wp-json\/wp\/v2\/tags?post=652"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}