Advanced Structured Prediction

摘要: The goal of structured prediction is to build machine learning models that predict relational information itself has structure, such as being composed multiple interrelated parts. These models, which reflect prior knowledge, task-specific relations, and constraints, are used in fields including computer vision, speech recognition, natural language processing, computational biology. They can carry out tasks predicting a sentence, or segmenting an image into meaningful components. expressive powerful, but exact computation often intractable. A broad research effort recent years aimed at designing approximate inference procedures computationally efficient. This volume offers overview this order make the work accessible broader community. chapters, by leading researchers field, cover range topics, trends, linear programming relaxation approach, innovations probabilistic modeling, theoretical progress, resource-aware learning.Sebastian Nowozin Researcher Machine Learning Perception group (MLP) Microsoft Research, Cambridge, England. Peter V. Gehler Senior Perceiving Systems Max Planck Institute for Intelligent Systems, Tbingen, Germany. Jeremy Jancsary Research Scientist Nuance Communications, Vienna. Christoph H. Lampert Assistant Professor Science Technology Austria, where he heads Computer Vision Learning. Contributors Jonas Behr, Yutian Chen, Fernando De La Torre, Justin Domke, Gehler, Andrew E. Gelfand, Sbastien Gigure, Amir Globerson, Fred A. Hamprecht, Minh Hoai, Tommi Jaakkola, Jancsary, Joseph Keshet, Marius Kloft, Vladimir Kolmogorov, Lampert, Franois Laviolette, Xinghua Lou, Mario Marchand, Andr F. T. Martins, Ofer Meshi, Sebastian Nowozin, George Papandreou, Daniel Pra, Gunnar Rtsch, Amlie Rolland, Bogdan Savchynskyy, Stefan Schmidt, Thomas Schoenemann, Gabriele Schweikert, Ben Taskar, Sinisa Todorovic, Welling, David Weiss, Thom Werner, Alan Yuille, Stanislav ivn