Advanced Structured Prediction
The goal of structured prediction is to build machine learning models that predict relational information that itself has structure, such as being composed of multiple interrelated parts. These models, which reflect prior knowledge, task-specific relations, and constraints, are used in fields includ...
Saved in:
| Main Authors | , , , |
|---|---|
| Format | eBook Book |
| Language | English |
| Published |
Cambridge, MA
MIT Press
2014
The MIT Press |
| Edition | 1 |
| Series | Neural Information Processing series |
| Subjects | |
| Online Access | Get full text |
| ISBN | 0262028379 9780262028370 |
| DOI | 10.7551/mitpress/9969.001.0001 |
Cover
| Abstract | The goal of structured prediction is to build machine learning models that predict relational information that itself has structure, such as being composed of 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, and computational biology. They can carry out such tasks as predicting a natural language sentence, or segmenting an image into meaningful components. These models are expressive and powerful, but exact computation is often intractable. A broad research effort in recent years has aimed at designing structured prediction models and approximate inference and learning procedures that are computationally efficient. This volume offers an overview of this recent research in order to make the work accessible to a broader research community. The chapters, by leading researchers in the field, cover a range of topics, including research trends, the linear programming relaxation approach, innovations in probabilistic modeling, recent theoretical progress, and resource-aware learning. Sebastian Nowozin is a Researcher in the Machine Learning and Perception group (MLP) at Microsoft Research, Cambridge, England. Peter V. Gehler is a Senior Researcher in the Perceiving Systems group at the Max Planck Institute for Intelligent Systems, Tübingen, Germany. Jeremy Jancsary is a Senior Research Scientist at Nuance Communications, Vienna. Christoph H. Lampert is Assistant Professor at the Institute of Science and Technology Austria, where he heads a group for Computer Vision and Machine Learning. Contributors Jonas Behr, Yutian Chen, Fernando De La Torre, Justin Domke, Peter V. Gehler, Andrew E. Gelfand, Sébastien Giguère, Amir Globerson, Fred A. Hamprecht, Minh Hoai, Tommi Jaakkola, Jeremy Jancsary, Joseph Keshet, Marius Kloft, Vladimir Kolmogorov, Christoph H. Lampert, François Laviolette, Xinghua Lou, Mario Marchand, André F. T. Martins, Ofer Meshi, Sebastian Nowozin, George Papandreou, Daniel Pruša, Gunnar Rätsch, Amélie Rolland, Bogdan Savchynskyy, Stefan Schmidt, Thomas Schoenemann, Gabriele Schweikert, Ben Taskar, Sinisa Todorovic, Max Welling, David Weiss, Thomáš Werner, Alan Yuille, Stanislav Živný |
|---|---|
| AbstractList | The goal of structured prediction is to build machine learning models that predict relational information that itself has structure, such as being composed of 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, and computational biology. They can carry out such tasks as predicting a natural language sentence, or segmenting an image into meaningful components. These models are expressive and powerful, but exact computation is often intractable. A broad research effort in recent years has aimed at designing structured prediction models and approximate inference and learning procedures that are computationally efficient. This volume offers an overview of this recent research in order to make the work accessible to a broader research community. The chapters, by leading researchers in the field, cover a range of topics, including research trends, the linear programming relaxation approach, innovations in probabilistic modeling, recent theoretical progress, and resource-aware learning.Sebastian Nowozin is a Researcher in the Machine Learning and Perception group (MLP) at Microsoft Research, Cambridge, England. Peter V. Gehler is a Senior Researcher in the Perceiving Systems group at the Max Planck Institute for Intelligent Systems, Tübingen, Germany. Jeremy Jancsary is a Senior Research Scientist at Nuance Communications, Vienna. Christoph H. Lampert is Assistant Professor at the Institute of Science and Technology Austria, where he heads a group for Computer Vision and Machine Learning. Contributors Jonas Behr, Yutian Chen, Fernando De La Torre, Justin Domke, Peter V. Gehler, Andrew E. Gelfand, Sébastien Giguère, Amir Globerson, Fred A. Hamprecht, Minh Hoai, Tommi Jaakkola, Jeremy Jancsary, Joseph Keshet, Marius Kloft, Vladimir Kolmogorov, Christoph H. Lampert, François Laviolette, Xinghua Lou, Mario Marchand, André F. T. Martins, Ofer Meshi, Sebastian Nowozin, George Papandreou, Daniel Pruša, Gunnar Rätsch, Amélie Rolland, Bogdan Savchynskyy, Stefan Schmidt, Thomas Schoenemann, Gabriele Schweikert, Ben Taskar, Sinisa Todorovic, Max Welling, David Weiss, Thomáš Werner, Alan Yuille, Stanislav Živný The goal of structured prediction is to build machine learning models that predict relational information that itself has structure, such as being composed of 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, and computational biology. They can carry out such tasks as predicting a natural language sentence, or segmenting an image into meaningful components. These models are expressive and powerful, but exact computation is often intractable. A broad research effort in recent years has aimed at designing structured prediction models and approximate inference and learning procedures that are computationally efficient. This volume offers an overview of this recent research in order to make the work accessible to a broader research community. The chapters, by leading researchers in the field, cover a range of topics, including research trends, the linear programming relaxation approach, innovations in probabilistic modeling, recent theoretical progress, and resource-aware learning.Sebastian Nowozin is a Researcher in the Machine Learning and Perception group (MLP) at Microsoft Research, Cambridge, England. Peter V. Gehler is a Senior Researcher in the Perceiving Systems group at the Max Planck Institute for Intelligent Systems, Tübingen, Germany. Jeremy Jancsary is a Senior Research Scientist at Nuance Communications, Vienna. Christoph H. Lampert is Assistant Professor at the Institute of Science and Technology Austria, where he heads a group for Computer Vision and Machine Learning. Contributors Jonas Behr, Yutian Chen, Fernando De La Torre, Justin Domke, Peter V. Gehler, Andrew E. Gelfand, Sébastien Giguère, Amir Globerson, Fred A. Hamprecht, Minh Hoai, Tommi Jaakkola, Jeremy Jancsary, Joseph Keshet, Marius Kloft, Vladimir Kolmogorov, Christoph H. Lampert, François Laviolette, Xinghua Lou, Mario Marchand, André F. T. Martins, Ofer Meshi, Sebastian Nowozin, George Papandreou, Daniel Pruša, Gunnar Rätsch, Amélie Rolland, Bogdan Savchynskyy, Stefan Schmidt, Thomas Schoenemann, Gabriele Schweikert, Ben Taskar, Sinisa Todorovic, Max Welling, David Weiss, Thomáš Werner, Alan Yuille, Stanislav Živný |
| Author | Jancsary, Jeremy Gehler, Peter V. Nowozin, Sebastian Lampert, Christoph H. |
| Author_xml | – sequence: 1 fullname: Nowozin, Sebastian – sequence: 2 fullname: Gehler, Peter V. – sequence: 3 fullname: Jancsary, Jeremy – sequence: 4 fullname: Lampert, Christoph H. |
| BackLink | https://cir.nii.ac.jp/crid/1130282272914843392$$DView record in CiNii |
| BookMark | eNqNkEtPwkAQx9coRkA-gBfiwcR4AHZn9tE5IsFHQqKJxmuz3W5jBVvtFvz6LtaDRw_zSn7zn8eAHVV15RkbCz41SonZe9l-ND6EGZGmKeciGhcHbEQm4aABAUjRIRvsCw4JGuqxAXAhuUBAdcz6xJUkRE4nbBTC275fKlQS-2w8z3e2cj4_f2qbrWu3TUwfoytdW9bVKesVdhP86DcO2cvN8nlxN1k93N4v5quJBS0AJ1Z6Cw4w15occulzU3DtnSukyZ3JMqLMF8JoMAp5AqTzQmY2E0USl_YCh-yqE7Zh7b_Ca71pQ7rb-Kyu1yH9c6rW_2cVRfayYz-a-nPrQ5v-YM5XbWM36fJ6oaKkUiqSFx1ZlWXqyr0XAuNDAQyQkIlEJIjYWYeV3vu0G2k4T4Tm-A2vz3fe |
| ContentType | eBook Book |
| Copyright | 2014 Massachusetts Institute of Technology |
| Copyright_xml | – notice: 2014 Massachusetts Institute of Technology |
| DBID | RYH |
| DEWEY | 006.3/1 |
| DOI | 10.7551/mitpress/9969.001.0001 |
| DatabaseName | CiNii Complete |
| DatabaseTitleList | |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Computer Science |
| EISBN | 9780262322959 0262322951 9780262322966 026232296X |
| Edition | 1 |
| Editor | Gehler, Peter V Jancsary, Jeremy Nowozin, Sebastian Lampert, Christoph H |
| Editor_xml | – sequence: 1 givenname: Sebastian surname: Nowozin fullname: Nowozin, Sebastian – sequence: 2 givenname: Peter V surname: Gehler fullname: Gehler, Peter V – sequence: 3 givenname: Jeremy surname: Jancsary fullname: Jancsary, Jeremy – sequence: 4 givenname: Christoph H surname: Lampert fullname: Lampert, Christoph H |
| ExternalDocumentID | 9780262322966 9780262322959 EBC5966555 BB17474033 7008160 |
| GroupedDBID | -D2 38. 6IK AABBV AAOBU ABAZT ABFEK AEFEZ AEGYG AGMVS AHWGJ ALMA_UNASSIGNED_HOLDINGS BBABE BEFXN BFFAM BGNUA BKEBE BPEOZ EBRZX ECNEQ MICIX MIJRL OCL RYH |
| ID | FETCH-LOGICAL-a26123-a4ea2c23d669c304ed7f06eccf47dc7bb99bef176275308296df4bab1f8620e13 |
| IEDL.DBID | -D2 |
| ISBN | 0262028379 9780262028370 |
| IngestDate | Wed Feb 05 01:58:43 EST 2025 Thu Mar 20 12:07:31 EDT 2025 Wed Jun 04 00:00:55 EDT 2025 Thu Jun 26 23:18:03 EDT 2025 Sun Apr 21 09:07:52 EDT 2024 |
| IsPeerReviewed | false |
| IsScholarly | false |
| LCCN | 2014013235 |
| LCCallNum_Ident | Q325.5$b.N696 2014 |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-a26123-a4ea2c23d669c304ed7f06eccf47dc7bb99bef176275308296df4bab1f8620e13 |
| Notes | Includes bibliographical references Other editors : Peter V. Gehler, Jeremy Jancsary, Christoph H. Lampert |
| OCLC | 905493309 |
| PQID | EBC5966555 |
| PageCount | 432 |
| ParticipantIDs | askewsholts_vlebooks_9780262322966 askewsholts_vlebooks_9780262322959 proquest_ebookcentral_EBC5966555 nii_cinii_1130282272914843392 ieee_books_7008160 |
| PublicationCentury | 2000 |
| PublicationDate | 2014 c2014 2015-01-16 2014-11-21 |
| PublicationDateYYYYMMDD | 2014-01-01 2015-01-16 2014-11-21 |
| PublicationDate_xml | – year: 2014 text: 2014 |
| PublicationDecade | 2010 |
| PublicationPlace | Cambridge, MA |
| PublicationPlace_xml | – name: Cambridge, MA – name: Cambridge |
| PublicationSeriesTitle | Neural Information Processing series |
| PublicationYear | 2014 2015 |
| Publisher | MIT Press The MIT Press |
| Publisher_xml | – name: MIT Press – name: The MIT Press |
| SSID | ssj0001453543 |
| Score | 1.9324222 |
| Snippet | The goal of structured prediction is to build machine learning models that predict relational information that itself has structure, such as being composed of... The goal of structured prediction is to build machine learning models that predict relational information that itself has structure, such as being composed of... |
| SourceID | askewsholts proquest nii ieee |
| SourceType | Aggregation Database Publisher |
| SubjectTerms | Computer algorithms Computing and Processing Data structures (Computer science) General Topics for Engineers Machine learning |
| TableOfContents | Intro -- Contents -- Series Foreword -- Preface -- 1. Introduction to Structured Prediction -- 1.1 Structured Prediction -- 1.2 Recent Developments -- 1.3 Summary of the Chapters -- 1.4 Conclusion -- 1.5 References -- 2. The Power of LP Relaxation for MAP Inference -- 2.1 Valued Constraint Satisfaction Problem -- 2.2 Basic LP Relaxation -- 2.3 Languages Solved by the BLP -- 2.4 Universality of the BLP -- 2.5 Conclusions -- 2.6 References -- 3 AD3: A Fast Decoder for Structured Prediction -- 3.1 Introduction -- 3.2 Factor Graphs and MAP Decoding -- 3.3 Factor Graphs for NLP -- 3.4 LP-MAP Decoding -- 3.5 Alternating Directions Dual Decomposition -- 3.6 Local Subproblems in AD3 -- 3.7 Experiments -- 3.8 Related Work -- 3.9 Conclusions -- 3.10 References -- 4 Generalized Sequential Tree-Reweighted Message Passing -- 4.1 Introduction -- 4.2 Background and Notation -- 4.3 TRW-S Algorithm -- 4.4 Algorithm's Analysis -- 4.5 Experimental Results -- 4.6 Conclusions -- 4.7 Appendices -- 4.8 References -- 5 Smoothed Coordinate Descent for MAP Inference -- 5.1 Introduction -- 5.2 MAP and LP Relaxations -- 5.3 Coordinate Minimization Algorithms -- 5.4 Dual Convergence Rate Analysis -- 5.5 Primal Convergence -- 5.6 The Augmented Dual LP Algorithm -- 5.7 Experiments -- 5.8 Discussion -- 5.9 Appendix: Primal Convergence Rate -- 5.10 References -- 6 Getting Feasible Variable Estimates from Infeasible Ones: MRF Local Polytope Study -- 6.1 Introduction -- 6.2 Optimizing Projection -- 6.3 MRF Inference and Optimizing Projections -- 6.4 Optimizing Projection in Algorithmic Schemes -- 6.5 Experimental Analysis and Evaluation -- 6.6 Conclusions -- 6.7 References -- 7 Perturb-and-MAP Random Fields: Reducing Random Sampling to Optimization, with Applications in Computer Vision -- 7.1 Introduction 7.2 Energy-Based Modeling: Standard Deterministic and Probabilistic Approaches -- 7.3 Perturb-and-MAP for Gaussian and Sparse Continuous MRFs -- 7.4 Perturb-and-MAP for MRFs with Discrete Labels -- 7.5 Related Work and Recent Developments -- 7.6 Discussion -- 7.7 References -- 8 Herding for Structured Prediction -- 8.1 Introduction -- 8.2 Integrating Local Models Using Herding -- 8.3 Application: Image Segmentation -- 8.4 Application: Go Game Prediction -- 8.5 Conclusion -- 8.6 References -- 9 Training Structured Predictors Through Iterated Logistic Regression -- 9.1 Introduction -- 9.2 Linear vs. Nonlinear Learning -- 9.3 Overview -- 9.4 Loss Functions -- 9.5 Message-Passing Inference -- 9.6 Joint Learning and Inference -- 9.7 Logistic Regression -- 9.8 Reducing Structured Learning to Logistic Regression -- 9.9 Function Classes -- 9.10 Example -- 9.11 Conclusions -- 9.12 Appendix: Proofs -- 9.13 References -- 10 PAC-Bayesian Risk Bounds and Learning Algorithms for the Regression Approach to Structured Output Prediction -- 10.1 Introduction -- 10.2 From Structured Output Prediction to Vector-Valued Regression -- 10.3 A PAC-Bayesian Bound with Isotropic Gaussians -- 10.4 A Sample Compressed PAC-Bayesian Bound -- 10.5 Empirical Results -- 10.6 Conclusion -- 10.7 Appendix -- 10.8 References -- 11 Optimizing the Measure of Performance -- 11.1 Introduction -- 11.2 Structured Perceptron -- 11.3 Large Margin Structured Predictors -- 11.4 Conditional Random Fields -- 11.5 Direct Loss Minimization -- 11.6 Structured Ramp Loss -- 11.7 Structured Probit Loss -- 11.8 Risk Minimization Under Gibbs Distribution -- 11.9 Conclusions -- 11.10 References -- 12 Structured Learning from Cheap Data -- 12.1 Introduction -- 12.2 Running Example: Structured Learning for Cell Tracking -- 12.3 Strategy I: Structured Learning from Partial Annotations 12.4 Strategy II: Structured Data Retrieval via Active Learning -- 12.5 Strategy III: Structured Transfer Learning -- 12.6 Discussion and Conclusions -- 12.7 References -- 13 Dynamic Structured Model Selection -- 13.1 Introduction -- 13.2 Meta-Learning a Myopic Value-Based Selector -- 13.3 Applications to Sequential Prediction -- 13.4 Meta-Learning a Feature Extraction Policy -- 13.5 Applications to Sequential Prediction Revisited -- 13.6 Conclusion -- 13.7 References -- 14 Structured Prediction for Event Detection -- 14.1 Introduction -- 14.2 Structured Prediction for Event Detection -- 14.3 Early Event Detection -- 14.4 Sequence Labeling -- 14.5 Experiments -- 14.6 Summary -- 14.7 References -- 15 Structured Prediction for Object Boundary Detection in Images -- 15.1 Introduction -- 15.2 Related Work -- 15.3 Edge Extraction and Properties -- 15.4 Sequential Labeling of Edges -- 15.5 HC-Search -- 15.6 Results -- 15.7 Conclusion -- 15.8 References -- 16 Genome Annotation with Structured Output Learning -- 16.1 Introduction: The Genome Annotation Problem -- 16.2 Inference -- 16.3 Learning -- 16.4 Experiments -- 16.5 Conclusions -- 16.6 References |
| Title | Advanced Structured Prediction |
| URI | https://ieeexplore.ieee.org/servlet/opac?bknumber=7008160 https://cir.nii.ac.jp/crid/1130282272914843392 https://ebookcentral.proquest.com/lib/[SITE_ID]/detail.action?docID=5966555 https://www.vlebooks.com/vleweb/product/openreader?id=none&isbn=9780262322959&uid=none https://www.vlebooks.com/vleweb/product/openreader?id=none&isbn=9780262322966&uid=none |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV3NS8MwFA-yXdxFh4pVN4Z47dY2H21OotMxPIjgB8NLSdIEynDC1l38630v7XCoB_FSAg2heS_Ne7-Xl98j5CLLtEsi6-BHEjZkikehMpkOtUksVULGxvpsi3sxfWZ3Mz7bKvWFFU588pkdYtOf5WN4EibhUeSlnjelMlK0ZgLQehsQDMXFHd4kX-EVxilntL4TnIJbMHorK59aOgIP33NUInVh3CEdtZrDfgJ7TbVqSqyAnVmU5Y_d2ZucyR553XxsnWkyH64rPTQf33gc_zWbfdK2eMmhS3bs4oD0r5pEgMGjZ5NdL6H5sMQjHFTbIXmZ3D6Np2FTNyFUSAhGQ8WsSkxCCyGkoRGzReoiAcpyLC1MqrWU2ro4RYZipKuRonBMKx07wDeRjekRaS3eF_aYDFwilU6cplZyBtZeSp4WNs24NNA7YgE535JfDjNDcLDKkdgoAdcKq4XLP3QSIiBdlFFev2rkEZAeaCI3JT5jPGQFhwbwAIA4RsG1C8hgo6PcD9qktOa312MOg3LOT34b95TsgufD6ljKGWmBdG0PvItK9_06-gQJe8qp |
| linkProvider | IEEE |
| linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwjV1LSwMxEA5SD9qLFhXrs4jXbXc3j22OPlrqqwhW6W1JshMo1Qpu9eCvdyZdUdSDlyWwISwz2cw3mZlvGDvudq1PY_D4IymIhJFxZFzXRtalwI3SiYOQbTFUg3txOZbj6r4j1MIAQEg-gzYNQyyf4GYnI-Ol0DlfRpsvaC9H5-nXbYqQXAq-KAHOEAV0nibzkEnaQUAfKCmJqTCps7opp3h84NEyL6uOKmhWZpPJr8M4WJj-2oLuqAzEhJRYMm2_zm3bvf-gbfzPx6-zZaAShgZbgtkGOzipwvytu8AV-_qCw9sXCtCQUjbZQ783OhtEVVeEyBDdF4-MAJO6lBdKacdjAUXmY4Wq8CIrXGat1hZ8khH_MJHRaFV4YY1NPHovMSR8i9VmzzPYZi2famNTbzloKdCWay2zArKu1A5nx6LJjr6JK397DBHcMifaohSBE_UC1_-YpFSTNUgk-eJVJY8m20fB525Cz4RCqAhXEO2jiyY4Arcma32qJA-LVgmree_0TOKiUsqdv9Y9ZCuD0c11fn0xvNplq4hxxOLWZI_VUNKwjzhibg_CFvoAzNy_2g |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=book&rft.title=Advanced+Structured+Prediction&rft.series=Neural+Information+Processing+series&rft.date=2014-01-01&rft.pub=MIT+Press&rft.isbn=9780262322959&rft_id=info:doi/10.7551%2Fmitpress%2F9969.001.0001&rft.externalDocID=7008160 |
| thumbnail_m | http://utb.summon.serialssolutions.com/2.0.0/image/custom?url=https%3A%2F%2Fvle.dmmserver.com%2Fmedia%2F640%2F97802623%2F9780262322959.jpg http://utb.summon.serialssolutions.com/2.0.0/image/custom?url=https%3A%2F%2Fvle.dmmserver.com%2Fmedia%2F640%2F97802623%2F9780262322966.jpg |