Call for Papers: Workshop on Method of Moments and Spectral Learning -- ICML 2014 June 25 or 26, Beijing (China) Website: https://sites.google.com/site/momentsicml2014/
Many problems in machine learning involve collecting high-dimensional multivariate observations or sequences of observations, and then fitting a compact model which explains these observations. Because of their flexibility and expressive power, models with latent variables are typically considered. The predominant approaches used in machine learning for fitting these models are based either on the principle of maximum likelihood or Bayesian inference. However, the algorithms used with these approaches (e.g., Expectation-Maximization) are known to suffer from slow convergence or poor quality local optima. In the past several years, the machine learning and computer science communities have revisited a classical statistical approach called the method of moments, and designed computationally efficient algorithms based on this approach to tackle challenging learning problems. Many of these algorithms have been based on spectral decompositions of moment matrices or other algebraic structures, and hence have also gone by the name of "spectral learning" algorithms. In contrast to algorithms like E-M, these algorithms come with polynomial computational and sample complexity guarantees. Moreover, they have been applied to learn the structure and parameters of many models including predictive state representations, finite state transducers, hidden Markov models, latent trees, latent junction trees, probabilistic context free grammars, and mixture/admixture models. They have also been applied to a wide range of application domains including system identification, video modeling, speech modeling, robotics, and natural language processing. The focus of this workshop will be on spectral learning algorithms and the application of the method of moments to machine learning problems. We would like the workshop to be as inclusive as possible and encourage paper submissions and participation from a wide range of research related to this focus. The workshop will feature invited talks from renowned researchers and a poster session with contributed papers. ** Invited Speakers ** - Anima Anandkumar (UC Irvine) - Elina Robeva (UC Berkeley) - Sujay Sanghavi (UT Austin) - Aravindan Vijayaraghavan (Carnegie Mellon University) ** Submissions ** Extended abstracts should be submitted using the ICML 2014 format with a maximum of 4 pages (not including references). Please e-mail your submission to <workshop.spectral.learn...@gmail.com> with the subject line "Submission to ICML Workshop". Concurrent submissions to the workshop and the main conference (or other conferences) are permitted. ** Important dates ** - Submission deadline (extended): April 10, 2014 - Notification of acceptance: April 20, 2014 - Workshop: June 25 or 26, 2014 ** Organizers ** - Borja Balle (McGill University) - Byron Boots (University of Washington) - Yoni Halpern (New York University) - Daniel Hsu (Columbia University) - Percy Liang (Stanford University) - David Sontag (New York University)
_______________________________________________ uai mailing list uai@ENGR.ORST.EDU https://secure.engr.oregonstate.edu/mailman/listinfo/uai
