Google Tech Talks February, 28 2008
Added: 6 months ago
Views: 1,491ABSTRACT
Treebank parsing can be seen as the search for an optimally refined grammar consistent with a coarse training treebank. We describe a method in which a minimal grammar is hierarchically refined using EM to give accurate, compact grammars. The resulting grammars are extremely compact compared to other high-performance parsers, yet the parser gives the best published accuracies on several languages, as well as the best generative parsing numbers in English. In addition, we give an associated coarse-to-fine inference scheme which vastly improves inference time with no loss in test set accuracy.
Slides: http://www.eecs.berkeley.edu/~petrov/...
Speaker: Slav Petrov
Slav Petrov is a Ph.D. Candidate at University of California Berkeley Dept of Computer Science, where he is also a research assistant working with Dan Klein and Jitendra Malik on inducing latent structure for perception problems in vision and language.
Treebank parsing can be seen as the search for an optimally refined grammar consistent with a coarse training treebank. We describe a method in which a minimal grammar is hierarchically refined using EM to give accurate, compact grammars. The resulting grammars are extremely compact compared to other high-performance parsers, yet the parser gives the best published accuracies on several languages, as well as the best generative parsing numbers in English. In addition, we give an associated coarse-to-fine inference scheme which vastly improves inference time with no loss in test set accuracy.
Slides: http://www.eecs.berkeley.edu/~petrov/...
Speaker: Slav Petrov
Slav Petrov is a Ph.D. Candidate at University of California Berkeley Dept of Computer Science, where he is also a research assistant working with Dan Klein and Jitendra Malik on inducing latent structure for perception problems in vision and language.
Google Tech Talks April, 17 2008 ABSTRACT Modeling human sentence-processing can help us (more)
Added: 4 months ago
Views: 7,892Python
01:40:15 googletechtalks
01:06:41 From: googletechtalks
Cognitive Science
Google Tech Talks November, 15 2007
Added: 8 months ago
Views: 9,573Miscellaneous
Google Tech Talks January, 29 2008 ABSTRACT IPv6 and the DNS Speaker: Suzanne Woolf (more)
June 4, 2008
Google Tech Talks June 4, 2008 ABSTRACT In software engineering, aspects are concerns t (more)
Added: 3 months ago
Views: 2,535ABSTRACT
In software engineering, aspects are concerns that cut across multiple modules. They can lead to the common problems of concern tangling and scattering: concern tangling is where software concerns are not represented independently of each other; concern scattering is where a software concern is represented in multiple remote places in a software artifact. Although aspect-oriented programming is relatively well understood, aspect-oriented modeling (i.e., the representation of aspects during requirements engineering, architecture, design) is still rather immature. Although a wide variety of approaches to aspect-oriented modeling have been suggested, there is, as yet, no common consensus on how aspect-oriented models should be captured, manipulated and reasoned about. This talk presents MATA (Modeling Aspects Using a Transformation Approach), which is a unified way of handling aspects for any well-defined modeling language. The talk will argue why MATA is necessary and highlight some of the key features of MATA. In particular, the talk will motivate the decision to base MATA on graph transformations and will describe an application of MATA to modeling security concerns.
Speaker: Jon Whittle
Prof. Jon Whittle joined Lancaster University in August 2007 as a Professor of Software Engineering. Previously, he was an Associate Professor at George Mason University, Fairfax, VA, USA, and, prior to that, he was a researcher and contractor technical area lead at NASA Ames Research Center. In July 2007, he was awarded a highly prestigious Wolfson Merit Award from the Royal Society in the UK. Jon's research interests are in model-driven software development, formal methods, secure software development, requirements engineering and domain-specific methods for software engineering. His research has been recognized by a number of Best Paper awards, including the IEE Software Premium prize (with João Araújo). He is Chair of the Steering
Committee of the International Conference on Model-Driven Engineering, Languages and Systems
and has been a program committee member of this conference since 2002 (including experience track PC chair in 2006). He has served on over 30 program committees for international conferences and workshops.
He is an Associate Editor of the Journal of Software and Systems Modeling. Jon has also been a guest editor of the IEEE Transactions on Software Engineering, the Journal of Software Quality, and has co-edited two special issues of the Journal of Software and Systems Modeling.
In software engineering, aspects are concerns that cut across multiple modules. They can lead to the common problems of concern tangling and scattering: concern tangling is where software concerns are not represented independently of each other; concern scattering is where a software concern is represented in multiple remote places in a software artifact. Although aspect-oriented programming is relatively well understood, aspect-oriented modeling (i.e., the representation of aspects during requirements engineering, architecture, design) is still rather immature. Although a wide variety of approaches to aspect-oriented modeling have been suggested, there is, as yet, no common consensus on how aspect-oriented models should be captured, manipulated and reasoned about. This talk presents MATA (Modeling Aspects Using a Transformation Approach), which is a unified way of handling aspects for any well-defined modeling language. The talk will argue why MATA is necessary and highlight some of the key features of MATA. In particular, the talk will motivate the decision to base MATA on graph transformations and will describe an application of MATA to modeling security concerns.
Speaker: Jon Whittle
Prof. Jon Whittle joined Lancaster University in August 2007 as a Professor of Software Engineering. Previously, he was an Associate Professor at George Mason University, Fairfax, VA, USA, and, prior to that, he was a researcher and contractor technical area lead at NASA Ames Research Center. In July 2007, he was awarded a highly prestigious Wolfson Merit Award from the Royal Society in the UK. Jon's research interests are in model-driven software development, formal methods, secure software development, requirements engineering and domain-specific methods for software engineering. His research has been recognized by a number of Best Paper awards, including the IEE Software Premium prize (with João Araújo). He is Chair of the Steering
Committee of the International Conference on Model-Driven Engineering, Languages and Systems
and has been a program committee member of this conference since 2002 (including experience track PC chair in 2006). He has served on over 30 program committees for international conferences and workshops.
He is an Associate Editor of the Journal of Software and Systems Modeling. Jon has also been a guest editor of the IEEE Transactions on Software Engineering, the Journal of Software Quality, and has co-edited two special issues of the Journal of Software and Systems Modeling.
browsed googletechtalks until 300 (WINE conf 2007)
No comments:
Post a Comment