- published: 17 Oct 2008
- views: 3752
Create your page here
-
remove the playlistAnaphora Resolution
-
remove the playlistLatest Videos
-
remove the playlistLongest Videos
- remove the playlistAnaphora Resolution
- remove the playlistLatest Videos
- remove the playlistLongest Videos
back to playlist
Google Tech Talks
October 15, 2008
ABSTRACT
EM (the Expectation Maximization algorithm) is a well known technique for unsupervised learning (where one does not have any hand labeled solutions available, but instead one must learn from the raw text). Unfortunately EM is known to fail to find good solutions in many (most?) applications on which it is tried. In this talk we present some recent work on using EM to learn how to resolve pronoun-anaphora: determining that "the dog" is the antecedent of "he" and "his" in "When Sally fed the dog he wagged his tail". For this application EM works strikingly well, determining tens of thousands of parameters and resulting in a program that produces state of the art results.
Speaker: Eugene Charniak
Eugene Charniak is University Professor of Computer Science at Brown University and past chair of the department. He received his A.B. degree in Physics from University of Chicago, and a Ph.D. from M.I.T. in Computer Science. He has published four books the most recent being Statistical Language Learning. He is a Fellow of the American Association of Artificial Intelligence and was previously a Councilor of the organization. His research has always been in the area of language understanding or technologies which relate to it. Over the last 15 years years he has been interested in statistical techniques for many areas of language processing including parsing, discourse and anaphora.
This video demonstrates how embodied experience is used to enable an AI to correctly resolve references in natural language (in this case, to tell what object an instance of the word "it" refers to).
The video was screen-captured from actual real-time interaction between a human and an AI in the Multiverse virtual world.
The virtual dog is controlled by the AI; the human avatar is controlled by a human player.
While a relatively simple example, it involves integration of a natural language engine (RelEx), a cognition engine (the OpenCog-based OpenPetBrain), and a virtual world engine (Multiverse).
The text box that sometimes shows at the bottom left of the video contains the text typed by the human player, to communicate with the AI dog. The thought bubbles and captions shown elsewhere in the video, were added to the video after it was recorded, to better explain what is going on.
For more information, please go to: http://www.opencog.org
- published: 23 Nov 2009
- views: 1105
Deep Dhillon, Chief Data Scientist at Alliance Health Networks, presents a talk titled "Mining Unstructured Healthcare Data" to computational linguistics students at the University of Washington on May 8, 2013.
Every day doctors, researchers and health care professionals publish their latest medical findings continuously adding to the world's formalized medical knowledge represented by a corpus of millions of peer reviewed research studies. Meanwhile, millions of patients suffering from various conditions, communicate with one another in online discussion forums across the web; they seek both social comfort and knowledge. Medify analyzes the unstructured text of these health care professionals and patients by performing a deep NLP based statistical and lexical rule based relation extraction ultimately culminating in a large, searchable index powering a rapidly growing site trafficked by doctors, health care professionals, and advanced patients. We discuss the system at a high level, demonstrate key functionality, and explore what it means to develop a system like this in the confines of a start up. In addition, we dive into details like ground truth gathering, efficacy assessment, model approaches, feature engineering, anaphora resolution and more.
- published: 16 May 2013
- views: 2393
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation.
Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7
- published: 04 Dec 2012
- views: 85
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation.
Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7
- published: 04 Dec 2012
- views: 252
A brief explanation of the literary devices anaphora and epistrophe featuring Charles Dickens and John Bender
- published: 13 Apr 2014
- views: 2377
Dr. Christopher Hanlon and Dr. C.C. Wharram discuss anaphora for the Close Reading Cooperative, the podcast in literary analysis for English majors. To subscribe to the podcast and receive weekly installments, visit our instruction page at http://www.eiu.edu/~english/podcast.php
- published: 03 Jun 2011
- views: 2755
Modeling the Lifespan of Discourse Entities with Application to Coreference Resolution
Dr. Marie-Catherine de Marneffe
The Ohio State University
Dec 5, Fri, 2014
This talk aims at providing a compelling illustration from the domain of discourse of how efforts in theoretical linguistics and natural language processing can complement each other, both for developing models and for assessing them in engineering contexts. A discourse typically involves numerous entities, but few are mentioned more than once. Distinguishing entities that die out after just one mention (singletons) from those that lead longer lives (coreferents) would dramatically simplify the hypothesis space for coreference resolution models, leading to increased performance. To realize these gains, we synthesize linguistic insights about the factors affecting discourse entity lifespans (especially negation, modality, and attitude predication) with existing results about the benefits of "surface" (part-of-speech and n-gram-based) features for coreference resolution and build a classifier for predicting the singleton/coreferent distinction. The model is effective in its own right. It gives quantitative evidence for linguistic theories, and the feature representations help to identify the anchor phrases in bridging anaphora as well. Furthermore, incorporating the model into two very different state-of-the-art coreference resolution systems, one rule-based and the other learning-based, yields significant performance improvements.
Bio
Marie-Catherine de Marneffe is an assistant professor in Linguistics at The Ohio State University. She received her PhD from Stanford University in December 2012 under the supervision of Christopher D. Manning. She is developing computational linguistic methods that capture what is conveyed by speakers beyond the literal meaning of the words they say. Primarily she wants to ground meanings in corpus data, and show how such meanings can drive pragmatic inference. She has also worked on Recognizing Textual Entailment and contributed to defining the Stanford Dependencies and the Universal Dependencies representations, which are practical representations of grammatical relations and predicate argument structure.
Speaker webpage: http://www.ling.ohio-state.edu/~mcdm/
- published: 08 Dec 2014
- views: 143
In linguistics, an anaphora (
/əˈnæfərə/) is a type of expression whose reference depends on another referential element. E.g., in the sentence 'Sally preferred the company of herself', 'herself' is an anaphoric expression in that it is coreferential with the expression in subject position. Usually, an anaphoric expression is represented by a pro-form or some other kind of deictic—for instance, a pronoun referring to its antecedent. The term anaphor, an English singular variant, is sometimes used to designate an individual use: "an anaphor is a linguistic entity which indicates a referential tie to some other linguistic entity in the same text."
Anaphora is an important concept for different reasons and on different levels. First, anaphora indicates "how discourse is constructed and maintained". Second, on the level of the sentence, anaphora binds different syntactical elements together. Third, in computational linguistics anaphora presents a challenge to natural language processing, since the identification of the reference can be challenging. Fourth, anaphora "tells us some things about how language is understood and processed", which is relevant to fields of linguistics interested in cognitive psychology.
This page contains text from Wikipedia, the Free Encyclopedia -
http://en.wikipedia.org/wiki/Anaphora_(linguistics)
This article is licensed under the Creative Commons Attribution-ShareAlike 3.0 Unported License, which means that you can copy and modify it as long as the entire work (including additions) remains under this license.
This article is licensed under the Creative Commons Attribution-ShareAlike 3.0 Unported License, which means that you can copy and modify it as long as the entire work (including additions) remains under this license.
- Loading...
-
72:53EM Works for Pronoun-Anaphora Resolution
EM Works for Pronoun-Anaphora ResolutionEM Works for Pronoun-Anaphora Resolution
Google Tech Talks October 15, 2008 ABSTRACT EM (the Expectation Maximization algorithm) is a well known technique for unsupervised learning (where one does not have any hand labeled solutions available, but instead one must learn from the raw text). Unfortunately EM is known to fail to find good solutions in many (most?) applications on which it is tried. In this talk we present some recent work on using EM to learn how to resolve pronoun-anaphora: determining that "the dog" is the antecedent of "he" and "his" in "When Sally fed the dog he wagged his tail". For this application EM works strikingly well, determining tens of thousands of parameters and resulting in a program that produces state of the art results. Speaker: Eugene Charniak Eugene Charniak is University Professor of Computer Science at Brown University and past chair of the department. He received his A.B. degree in Physics from University of Chicago, and a Ph.D. from M.I.T. in Computer Science. He has published four books the most recent being Statistical Language Learning. He is a Fellow of the American Association of Artificial Intelligence and was previously a Councilor of the organization. His research has always been in the area of language understanding or technologies which relate to it. Over the last 15 years years he has been interested in statistical techniques for many areas of language processing including parsing, discourse and anaphora. -
5:06AnaphoraResolution1.mp4
AnaphoraResolution1.mp4 -
0:54Demo of Embodied Anaphor Resolution
Demo of Embodied Anaphor ResolutionDemo of Embodied Anaphor Resolution
This video demonstrates how embodied experience is used to enable an AI to correctly resolve references in natural language (in this case, to tell what object an instance of the word "it" refers to). The video was screen-captured from actual real-time interaction between a human and an AI in the Multiverse virtual world. The virtual dog is controlled by the AI; the human avatar is controlled by a human player. While a relatively simple example, it involves integration of a natural language engine (RelEx), a cognition engine (the OpenCog-based OpenPetBrain), and a virtual world engine (Multiverse). The text box that sometimes shows at the bottom left of the video contains the text typed by the human player, to communicate with the AI dog. The thought bubbles and captions shown elsewhere in the video, were added to the video after it was recorded, to better explain what is going on. For more information, please go to: http://www.opencog.org -
51:00Mining Unstructured Healthcare Data
Mining Unstructured Healthcare DataMining Unstructured Healthcare Data
Deep Dhillon, Chief Data Scientist at Alliance Health Networks, presents a talk titled "Mining Unstructured Healthcare Data" to computational linguistics students at the University of Washington on May 8, 2013. Every day doctors, researchers and health care professionals publish their latest medical findings continuously adding to the world's formalized medical knowledge represented by a corpus of millions of peer reviewed research studies. Meanwhile, millions of patients suffering from various conditions, communicate with one another in online discussion forums across the web; they seek both social comfort and knowledge. Medify analyzes the unstructured text of these health care professionals and patients by performing a deep NLP based statistical and lexical rule based relation extraction ultimately culminating in a large, searchable index powering a rapidly growing site trafficked by doctors, health care professionals, and advanced patients. We discuss the system at a high level, demonstrate key functionality, and explore what it means to develop a system like this in the confines of a start up. In addition, we dive into details like ground truth gathering, efficacy assessment, model approaches, feature engineering, anaphora resolution and more. -
15:16Digital Classicist Seminar Berlin (2012/2013) - Seminar 2 Discussion
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2 DiscussionDigital Classicist Seminar Berlin (2012/2013) - Seminar 2 Discussion
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project" Abstract: What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation. Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7 -
42:16Digital Classicist Seminar Berlin (2012/2013) - Seminar 2
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2Digital Classicist Seminar Berlin (2012/2013) - Seminar 2
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project" Abstract: What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation. Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7 -
2:33Anaphora and Epistrophe
Anaphora and EpistropheAnaphora and Epistrophe
A brief explanation of the literary devices anaphora and epistrophe featuring Charles Dickens and John Bender -
1:13Anaphora
AnaphoraAnaphora
For English -
5:34Anaphora
AnaphoraAnaphora
Dr. Christopher Hanlon and Dr. C.C. Wharram discuss anaphora for the Close Reading Cooperative, the podcast in literary analysis for English majors. To subscribe to the podcast and receive weekly installments, visit our instruction page at http://www.eiu.edu/~english/podcast.php -
51:22Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...
Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...
Modeling the Lifespan of Discourse Entities with Application to Coreference Resolution Dr. Marie-Catherine de Marneffe The Ohio State University Dec 5, Fri, 2014 This talk aims at providing a compelling illustration from the domain of discourse of how efforts in theoretical linguistics and natural language processing can complement each other, both for developing models and for assessing them in engineering contexts. A discourse typically involves numerous entities, but few are mentioned more than once. Distinguishing entities that die out after just one mention (singletons) from those that lead longer lives (coreferents) would dramatically simplify the hypothesis space for coreference resolution models, leading to increased performance. To realize these gains, we synthesize linguistic insights about the factors affecting discourse entity lifespans (especially negation, modality, and attitude predication) with existing results about the benefits of "surface" (part-of-speech and n-gram-based) features for coreference resolution and build a classifier for predicting the singleton/coreferent distinction. The model is effective in its own right. It gives quantitative evidence for linguistic theories, and the feature representations help to identify the anchor phrases in bridging anaphora as well. Furthermore, incorporating the model into two very different state-of-the-art coreference resolution systems, one rule-based and the other learning-based, yields significant performance improvements. Bio Marie-Catherine de Marneffe is an assistant professor in Linguistics at The Ohio State University. She received her PhD from Stanford University in December 2012 under the supervision of Christopher D. Manning. She is developing computational linguistic methods that capture what is conveyed by speakers beyond the literal meaning of the words they say. Primarily she wants to ground meanings in corpus data, and show how such meanings can drive pragmatic inference. She has also worked on Recognizing Textual Entailment and contributed to defining the Stanford Dependencies and the Universal Dependencies representations, which are practical representations of grammatical relations and predicate argument structure. Speaker webpage: http://www.ling.ohio-state.edu/~mcdm/
-
EM Works for Pronoun-Anaphora Resolution
Google Tech Talks October 15, 2008 ABSTRACT EM (the Expectation Maximization algorithm) is a well known technique for unsupervised learning (where one does not have any hand labeled solutions available, but instead one must learn from the raw text). Unfortunately EM is known to fail to find good solutions in many (most?) applications on which it is tried. In this talk we present some recent work on using EM to learn how to resolve pronoun-anaphora: determining that "the dog" is the antecedent of "he" and "his" in "When Sally fed the dog he wagged his tail". For this application EM works strikingly well, determining tens of thousands of parameters and resulting in a program that produces state of the art results. Speaker: Eugene Charniak Eugene Charniak is University Professor of Comput... -
-
Demo of Embodied Anaphor Resolution
This video demonstrates how embodied experience is used to enable an AI to correctly resolve references in natural language (in this case, to tell what object an instance of the word "it" refers to). The video was screen-captured from actual real-time interaction between a human and an AI in the Multiverse virtual world. The virtual dog is controlled by the AI; the human avatar is controlled by a human player. While a relatively simple example, it involves integration of a natural language engine (RelEx), a cognition engine (the OpenCog-based OpenPetBrain), and a virtual world engine (Multiverse). The text box that sometimes shows at the bottom left of the video contains the text typed by the human player, to communicate with the AI dog. The thought bubbles and captions shown e... -
Mining Unstructured Healthcare Data
Deep Dhillon, Chief Data Scientist at Alliance Health Networks, presents a talk titled "Mining Unstructured Healthcare Data" to computational linguistics students at the University of Washington on May 8, 2013. Every day doctors, researchers and health care professionals publish their latest medical findings continuously adding to the world's formalized medical knowledge represented by a corpus of millions of peer reviewed research studies. Meanwhile, millions of patients suffering from various conditions, communicate with one another in online discussion forums across the web; they seek both social comfort and knowledge. Medify analyzes the unstructured text of these health care professionals and patients by performing a deep NLP based statistical and lexical rule based relation extracti... -
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2 Discussion
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project" Abstract: What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (... -
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project" Abstract: What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (... -
Anaphora and Epistrophe
A brief explanation of the literary devices anaphora and epistrophe featuring Charles Dickens and John Bender -
Anaphora
For English -
Anaphora
Dr. Christopher Hanlon and Dr. C.C. Wharram discuss anaphora for the Close Reading Cooperative, the podcast in literary analysis for English majors. To subscribe to the podcast and receive weekly installments, visit our instruction page at http://www.eiu.edu/~english/podcast.php -
Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...
Modeling the Lifespan of Discourse Entities with Application to Coreference Resolution Dr. Marie-Catherine de Marneffe The Ohio State University Dec 5, Fri, 2014 This talk aims at providing a compelling illustration from the domain of discourse of how efforts in theoretical linguistics and natural language processing can complement each other, both for developing models and for assessing them in engineering contexts. A discourse typically involves numerous entities, but few are mentioned more than once. Distinguishing entities that die out after just one mention (singletons) from those that lead longer lives (coreferents) would dramatically simplify the hypothesis space for coreference resolution models, leading to increased performance. To realize these gains, we synthesize linguistic ...
EM Works for Pronoun-Anaphora Resolution
- Order: Reorder
- Duration: 72:53
- Updated: 17 Oct 2008
- views: 3752
Google Tech Talks
October 15, 2008
ABSTRACT
EM (the Expectation Maximization algorithm) is a well known technique for unsupervised learning (where one does no...
Google Tech Talks
October 15, 2008
ABSTRACT
EM (the Expectation Maximization algorithm) is a well known technique for unsupervised learning (where one does not have any hand labeled solutions available, but instead one must learn from the raw text). Unfortunately EM is known to fail to find good solutions in many (most?) applications on which it is tried. In this talk we present some recent work on using EM to learn how to resolve pronoun-anaphora: determining that "the dog" is the antecedent of "he" and "his" in "When Sally fed the dog he wagged his tail". For this application EM works strikingly well, determining tens of thousands of parameters and resulting in a program that produces state of the art results.
Speaker: Eugene Charniak
Eugene Charniak is University Professor of Computer Science at Brown University and past chair of the department. He received his A.B. degree in Physics from University of Chicago, and a Ph.D. from M.I.T. in Computer Science. He has published four books the most recent being Statistical Language Learning. He is a Fellow of the American Association of Artificial Intelligence and was previously a Councilor of the organization. His research has always been in the area of language understanding or technologies which relate to it. Over the last 15 years years he has been interested in statistical techniques for many areas of language processing including parsing, discourse and anaphora.
wn.com/Em Works For Pronoun Anaphora Resolution
Google Tech Talks
October 15, 2008
ABSTRACT
EM (the Expectation Maximization algorithm) is a well known technique for unsupervised learning (where one does not have any hand labeled solutions available, but instead one must learn from the raw text). Unfortunately EM is known to fail to find good solutions in many (most?) applications on which it is tried. In this talk we present some recent work on using EM to learn how to resolve pronoun-anaphora: determining that "the dog" is the antecedent of "he" and "his" in "When Sally fed the dog he wagged his tail". For this application EM works strikingly well, determining tens of thousands of parameters and resulting in a program that produces state of the art results.
Speaker: Eugene Charniak
Eugene Charniak is University Professor of Computer Science at Brown University and past chair of the department. He received his A.B. degree in Physics from University of Chicago, and a Ph.D. from M.I.T. in Computer Science. He has published four books the most recent being Statistical Language Learning. He is a Fellow of the American Association of Artificial Intelligence and was previously a Councilor of the organization. His research has always been in the area of language understanding or technologies which relate to it. Over the last 15 years years he has been interested in statistical techniques for many areas of language processing including parsing, discourse and anaphora.
- published: 17 Oct 2008
- views: 3752
AnaphoraResolution1.mp4
- Order: Reorder
- Duration: 5:06
- Updated: 30 Mar 2010
- views: 140
MediaHub Start-up and testing of Anaphora Resolution Scenario.
MediaHub Start-up and testing of Anaphora Resolution Scenario.
wn.com/Anaphoraresolution1.Mp4
Demo of Embodied Anaphor Resolution
- Order: Reorder
- Duration: 0:54
- Updated: 23 Nov 2009
- views: 1105
This video demonstrates how embodied experience is used to enable an AI to correctly resolve references in natural language (in this case, to tell what object a...
This video demonstrates how embodied experience is used to enable an AI to correctly resolve references in natural language (in this case, to tell what object an instance of the word "it" refers to).
The video was screen-captured from actual real-time interaction between a human and an AI in the Multiverse virtual world.
The virtual dog is controlled by the AI; the human avatar is controlled by a human player.
While a relatively simple example, it involves integration of a natural language engine (RelEx), a cognition engine (the OpenCog-based OpenPetBrain), and a virtual world engine (Multiverse).
The text box that sometimes shows at the bottom left of the video contains the text typed by the human player, to communicate with the AI dog. The thought bubbles and captions shown elsewhere in the video, were added to the video after it was recorded, to better explain what is going on.
For more information, please go to: http://www.opencog.org
wn.com/Demo Of Embodied Anaphor Resolution
This video demonstrates how embodied experience is used to enable an AI to correctly resolve references in natural language (in this case, to tell what object an instance of the word "it" refers to).
The video was screen-captured from actual real-time interaction between a human and an AI in the Multiverse virtual world.
The virtual dog is controlled by the AI; the human avatar is controlled by a human player.
While a relatively simple example, it involves integration of a natural language engine (RelEx), a cognition engine (the OpenCog-based OpenPetBrain), and a virtual world engine (Multiverse).
The text box that sometimes shows at the bottom left of the video contains the text typed by the human player, to communicate with the AI dog. The thought bubbles and captions shown elsewhere in the video, were added to the video after it was recorded, to better explain what is going on.
For more information, please go to: http://www.opencog.org
- published: 23 Nov 2009
- views: 1105
Mining Unstructured Healthcare Data
- Order: Reorder
- Duration: 51:00
- Updated: 16 May 2013
- views: 2393
Deep Dhillon, Chief Data Scientist at Alliance Health Networks, presents a talk titled "Mining Unstructured Healthcare Data" to computational linguistics studen...
Deep Dhillon, Chief Data Scientist at Alliance Health Networks, presents a talk titled "Mining Unstructured Healthcare Data" to computational linguistics students at the University of Washington on May 8, 2013.
Every day doctors, researchers and health care professionals publish their latest medical findings continuously adding to the world's formalized medical knowledge represented by a corpus of millions of peer reviewed research studies. Meanwhile, millions of patients suffering from various conditions, communicate with one another in online discussion forums across the web; they seek both social comfort and knowledge. Medify analyzes the unstructured text of these health care professionals and patients by performing a deep NLP based statistical and lexical rule based relation extraction ultimately culminating in a large, searchable index powering a rapidly growing site trafficked by doctors, health care professionals, and advanced patients. We discuss the system at a high level, demonstrate key functionality, and explore what it means to develop a system like this in the confines of a start up. In addition, we dive into details like ground truth gathering, efficacy assessment, model approaches, feature engineering, anaphora resolution and more.
wn.com/Mining Unstructured Healthcare Data
Deep Dhillon, Chief Data Scientist at Alliance Health Networks, presents a talk titled "Mining Unstructured Healthcare Data" to computational linguistics students at the University of Washington on May 8, 2013.
Every day doctors, researchers and health care professionals publish their latest medical findings continuously adding to the world's formalized medical knowledge represented by a corpus of millions of peer reviewed research studies. Meanwhile, millions of patients suffering from various conditions, communicate with one another in online discussion forums across the web; they seek both social comfort and knowledge. Medify analyzes the unstructured text of these health care professionals and patients by performing a deep NLP based statistical and lexical rule based relation extraction ultimately culminating in a large, searchable index powering a rapidly growing site trafficked by doctors, health care professionals, and advanced patients. We discuss the system at a high level, demonstrate key functionality, and explore what it means to develop a system like this in the confines of a start up. In addition, we dive into details like ground truth gathering, efficacy assessment, model approaches, feature engineering, anaphora resolution and more.
- published: 16 May 2013
- views: 2393
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2 Discussion
- Order: Reorder
- Duration: 15:16
- Updated: 04 Dec 2012
- views: 85
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribut...
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation.
Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7
wn.com/Digital Classicist Seminar Berlin (2012 2013) Seminar 2 Discussion
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation.
Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7
- published: 04 Dec 2012
- views: 85
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2
- Order: Reorder
- Duration: 42:16
- Updated: 04 Dec 2012
- views: 252
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribut...
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation.
Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7
wn.com/Digital Classicist Seminar Berlin (2012 2013) Seminar 2
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Linguistic annotation for the Hellespont Project"
Abstract:
What contribution can digital collections give to research in Ancient History? In order to answer this question, digital historians have rightly concentrated on the problem of how to structure the different documents (such as texts, surviving artifacts, geographical locations and related published works), so that the relevant materials can be retrieved with meaningful content-oriented queries. Among the projects that are pursuing this goal, the Hellespont Project (DAI, Perseus Project) focuses on the history of Athens in the years 479-431 BCE, as narrated in the text of Thucydides' Histories (I, 89-118). Yet, written sources for Ancient History (such as the works of Ancient historians) are especially complex. In order to get access to their content, they need structuring at a far more advanced level than current digital editions can provide. We propose to use the methods of current computational linguistics to address this issue. In particular we will explore how, in the Hellespont Project, we are taking advantage of the available annotated syntactic corpora and upgrading their model with supplementary annotation. Our goal is to enrich the text of Thucydides with with word-by-word linguistic annotation on morphology, syntax, valency frame and other discursive features such as semantic roles, verbal aspect, anaphora resolution and topic-focus articulation.
Further Information: http://hdl.handle.net/11858/00-1780-0000-000B-21BC-7
- published: 04 Dec 2012
- views: 252
Anaphora and Epistrophe
- Order: Reorder
- Duration: 2:33
- Updated: 13 Apr 2014
- views: 2377
A brief explanation of the literary devices anaphora and epistrophe featuring Charles Dickens and John Bender
A brief explanation of the literary devices anaphora and epistrophe featuring Charles Dickens and John Bender
wn.com/Anaphora And Epistrophe
A brief explanation of the literary devices anaphora and epistrophe featuring Charles Dickens and John Bender
- published: 13 Apr 2014
- views: 2377
Anaphora
- Order: Reorder
- Duration: 1:13
- Updated: 03 Nov 2010
- views: 2649
For English
For English
wn.com/Anaphora
Anaphora
- Order: Reorder
- Duration: 5:34
- Updated: 03 Jun 2011
- views: 2755
Dr. Christopher Hanlon and Dr. C.C. Wharram discuss anaphora for the Close Reading Cooperative, the podcast in literary analysis for English majors. To subscrib...
Dr. Christopher Hanlon and Dr. C.C. Wharram discuss anaphora for the Close Reading Cooperative, the podcast in literary analysis for English majors. To subscribe to the podcast and receive weekly installments, visit our instruction page at http://www.eiu.edu/~english/podcast.php
wn.com/Anaphora
Dr. Christopher Hanlon and Dr. C.C. Wharram discuss anaphora for the Close Reading Cooperative, the podcast in literary analysis for English majors. To subscribe to the podcast and receive weekly installments, visit our instruction page at http://www.eiu.edu/~english/podcast.php
- published: 03 Jun 2011
- views: 2755
Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...
- Order: Reorder
- Duration: 51:22
- Updated: 08 Dec 2014
- views: 143
Modeling the Lifespan of Discourse Entities with Application to Coreference Resolution
Dr. Marie-Catherine de Marneffe
The Ohio State University
Dec 5, Fri, 2...
Modeling the Lifespan of Discourse Entities with Application to Coreference Resolution
Dr. Marie-Catherine de Marneffe
The Ohio State University
Dec 5, Fri, 2014
This talk aims at providing a compelling illustration from the domain of discourse of how efforts in theoretical linguistics and natural language processing can complement each other, both for developing models and for assessing them in engineering contexts. A discourse typically involves numerous entities, but few are mentioned more than once. Distinguishing entities that die out after just one mention (singletons) from those that lead longer lives (coreferents) would dramatically simplify the hypothesis space for coreference resolution models, leading to increased performance. To realize these gains, we synthesize linguistic insights about the factors affecting discourse entity lifespans (especially negation, modality, and attitude predication) with existing results about the benefits of "surface" (part-of-speech and n-gram-based) features for coreference resolution and build a classifier for predicting the singleton/coreferent distinction. The model is effective in its own right. It gives quantitative evidence for linguistic theories, and the feature representations help to identify the anchor phrases in bridging anaphora as well. Furthermore, incorporating the model into two very different state-of-the-art coreference resolution systems, one rule-based and the other learning-based, yields significant performance improvements.
Bio
Marie-Catherine de Marneffe is an assistant professor in Linguistics at The Ohio State University. She received her PhD from Stanford University in December 2012 under the supervision of Christopher D. Manning. She is developing computational linguistic methods that capture what is conveyed by speakers beyond the literal meaning of the words they say. Primarily she wants to ground meanings in corpus data, and show how such meanings can drive pragmatic inference. She has also worked on Recognizing Textual Entailment and contributed to defining the Stanford Dependencies and the Universal Dependencies representations, which are practical representations of grammatical relations and predicate argument structure.
Speaker webpage: http://www.ling.ohio-state.edu/~mcdm/
wn.com/Dr. Marie Catherine De Marneffe Modeling The Lifespan Of Discourse Entities With Application To ...
Modeling the Lifespan of Discourse Entities with Application to Coreference Resolution
Dr. Marie-Catherine de Marneffe
The Ohio State University
Dec 5, Fri, 2014
This talk aims at providing a compelling illustration from the domain of discourse of how efforts in theoretical linguistics and natural language processing can complement each other, both for developing models and for assessing them in engineering contexts. A discourse typically involves numerous entities, but few are mentioned more than once. Distinguishing entities that die out after just one mention (singletons) from those that lead longer lives (coreferents) would dramatically simplify the hypothesis space for coreference resolution models, leading to increased performance. To realize these gains, we synthesize linguistic insights about the factors affecting discourse entity lifespans (especially negation, modality, and attitude predication) with existing results about the benefits of "surface" (part-of-speech and n-gram-based) features for coreference resolution and build a classifier for predicting the singleton/coreferent distinction. The model is effective in its own right. It gives quantitative evidence for linguistic theories, and the feature representations help to identify the anchor phrases in bridging anaphora as well. Furthermore, incorporating the model into two very different state-of-the-art coreference resolution systems, one rule-based and the other learning-based, yields significant performance improvements.
Bio
Marie-Catherine de Marneffe is an assistant professor in Linguistics at The Ohio State University. She received her PhD from Stanford University in December 2012 under the supervision of Christopher D. Manning. She is developing computational linguistic methods that capture what is conveyed by speakers beyond the literal meaning of the words they say. Primarily she wants to ground meanings in corpus data, and show how such meanings can drive pragmatic inference. She has also worked on Recognizing Textual Entailment and contributed to defining the Stanford Dependencies and the Universal Dependencies representations, which are practical representations of grammatical relations and predicate argument structure.
Speaker webpage: http://www.ling.ohio-state.edu/~mcdm/
- published: 08 Dec 2014
- views: 143
-
Truthseiyer's Minecraft Adventure Ep-015-pt2
Today I nearly finish my trap. The whole drag 'em in and kill 'em works fine, but I seem to be having trouble with the whole, "Get something out of it" part.
Truthseiyer's Minecraft Adventure Ep-015-pt2
- Order: Reorder
- Duration: 14:03
- Updated: 03 Apr 2011
- views: 28
Today I nearly finish my trap. The whole drag 'em in and kill 'em works fine, but I seem to be having trouble with the whole, "Get something out of it" part.
Today I nearly finish my trap. The whole drag 'em in and kill 'em works fine, but I seem to be having trouble with the whole, "Get something out of it" part.
wn.com/Truthseiyer's Minecraft Adventure Ep 015 Pt2
Today I nearly finish my trap. The whole drag 'em in and kill 'em works fine, but I seem to be having trouble with the whole, "Get something out of it" part.
- published: 03 Apr 2011
- views: 28
-
Active Learning for Clinical Coreference Resolution
Tutorial: Tim Miller Active Learning for Clinical Co-reference Resolution iDash NLP Annotation Workshop Saturday, September 29, 2012 Calit2, Atkinson Hall Auditorium, UCSD -
Coreference Resolution for Morphologically Rich ...
"Coreference Resolution for Morphologically Rich Languages. Adaptation of the Stanford System to Basque ".Ander Soraluze, Olatz Arregi, Xabier Arregi, Arantza Díaz de Ilarraza -
Scaling Language Understanding via Joint Multilingual Learning - Andreea Bodnari
See the full post here: Natural language is a pervasive human skill not yet fully achievable by automated computing systems. The main challenge is understanding how to computationally model both the depth and the breadth of natural languages. In this talk, Andreea Bodnari (Chief Data Scientist at Movable Ink) presents two probabilistic models that systematically model both the depth and the breadth of natural languages for two different linguistic tasks: syntactic parsing and joint learning of named entity recognition and coreference resolution. The syntactic parsing model outperforms current state of the art models by discovering linguistic information shared across languages at the granular level of a sentence. The coreference resolution system is one of the first attempts at joint mu... -
Professor Manfred Krifka - Research Seminar Series
Polarity Particles as Propositional Anaphors Recently, a number of theories have been proposed to explain polarity particles like yes and no (cf. Kramer & Rawlins 2009, Holmberg 2012, Farkas & Roelofsen 2012). I propose an analysis of polarity particles as anaphoric expressions that pick up propositional discourse referents, and that are themselves of the category of a clause. If the antecedent clauses contain negations, two such discourse referents are introduced, leading to the well-known ambiguities: John isn't coming to the party. -- No, he isn't, but also Yes, he isn't. It will be shown that pragmatic optimization leads to the selection of favourite answers in systems like English (with yes and no) and German (with ja, nein and doch). -
People Recognition - A Leapfrog in Organizing Videos
Google Tech Talks October 2, 2008 ABSTRACT Shifting from its initial purely entertainment-centric agenda, video is becoming an important media for communication, saving generic facts and personal memories. Viewdle is considering face recognition in real-life videos as one of the three material steps ahead (along with tracking location and timecodes) towards effective organization of videos. The presentation will focus on algorithmic and engineering approaches to address practical challenges of object recognition in low-resolution professional and user-generated videos. One of such approaches is multimodal fusion from several sources of information to achieve best-possible tagging and people-recognition-in-video experience. Viewdle outlines several key principles for "fusing" decisions fr... -
Let's Play Scribblenauts Unlimited [Deutsch] - Part 9 : Der Pandahaltige Metapher-Wald
Facebook ► http://tiny.cc/8t76ow Youtube ► http://tiny.cc/tu76ow - Informationen zum Game / LP - Spiel/Game : Scribblenauts Unlimited System/Platform : PC Entwickler/Developer : 5th Cell Media Publisher : Warner Bros. Interactive Entertainment Sprache/Language : German / Deutsch Spieler/Player + Kommentar/Commentary : Rivon Auflösung/Resolution : 1280x720 Genre : Puzzle, Action Release : 19. November 2012 Ihr habt jetz total lust auf diesen Titel bekommen und wollt ihn unbedingt auch mal zocken? Dann könnt ihr euch das Game hier bei Steam kaufen : Steam http://store.steampowered.com/app/218680 - Social Network und son Kram - Facebook ► http://tiny.cc/8t76ow Rivon Xfire ► http://tiny.cc/ov76ow Derfer Xfire ► http://tiny.cc/rw76ow Rivon Steam ► http://tiny.cc/cx76ow Derfer Steam ► ht...
Active Learning for Clinical Coreference Resolution
- Order: Reorder
- Duration: 27:46
- Updated: 02 Nov 2012
- views: 180
Tutorial: Tim Miller
Active Learning for Clinical Co-reference Resolution
iDash NLP Annotation Workshop
Saturday, September 29, 2012
Calit2, Atkinson Hall Aud...
Tutorial: Tim Miller
Active Learning for Clinical Co-reference Resolution
iDash NLP Annotation Workshop
Saturday, September 29, 2012
Calit2, Atkinson Hall Auditorium, UCSD
wn.com/Active Learning For Clinical Coreference Resolution
Tutorial: Tim Miller
Active Learning for Clinical Co-reference Resolution
iDash NLP Annotation Workshop
Saturday, September 29, 2012
Calit2, Atkinson Hall Auditorium, UCSD
- published: 02 Nov 2012
- views: 180
Coreference Resolution for Morphologically Rich ...
- Order: Reorder
- Duration: 29:41
- Updated: 22 Dec 2015
- views: 41
"Coreference Resolution for Morphologically Rich Languages. Adaptation of the Stanford System to Basque
".Ander Soraluze, Olatz Arregi, Xabier Arregi, Arantza ...
"Coreference Resolution for Morphologically Rich Languages. Adaptation of the Stanford System to Basque
".Ander Soraluze, Olatz Arregi, Xabier Arregi, Arantza Díaz de Ilarraza
wn.com/Coreference Resolution For Morphologically Rich ...
"Coreference Resolution for Morphologically Rich Languages. Adaptation of the Stanford System to Basque
".Ander Soraluze, Olatz Arregi, Xabier Arregi, Arantza Díaz de Ilarraza
- published: 22 Dec 2015
- views: 41
Scaling Language Understanding via Joint Multilingual Learning - Andreea Bodnari
- Order: Reorder
- Duration: 47:37
- Updated: 13 Nov 2014
- views: 362
See the full post here:
Natural language is a pervasive human skill not yet fully achievable by automated computing systems. The main challenge is understandin...
See the full post here:
Natural language is a pervasive human skill not yet fully achievable by automated computing systems. The main challenge is understanding how to computationally model both the depth and the breadth of natural languages.
In this talk, Andreea Bodnari (Chief Data Scientist at Movable Ink) presents two probabilistic models that systematically model both the depth and the breadth of natural languages for two different linguistic tasks: syntactic parsing and joint learning of named entity recognition and coreference resolution.
The syntactic parsing model outperforms current state of the art models by discovering linguistic information shared across languages at the granular level of a sentence. The coreference resolution system is one of the first attempts at joint multilingual modeling of named entity recognition and coreference resolution with limited linguistic resources. It performs second best on three out of four languages when compared to state of the art systems built with rich linguistic resources. Andreea shows that we can simultaneously model both the depth and the breadth of natural languages using the underlying linguistic structure shared across languages.
wn.com/Scaling Language Understanding Via Joint Multilingual Learning Andreea Bodnari
See the full post here:
Natural language is a pervasive human skill not yet fully achievable by automated computing systems. The main challenge is understanding how to computationally model both the depth and the breadth of natural languages.
In this talk, Andreea Bodnari (Chief Data Scientist at Movable Ink) presents two probabilistic models that systematically model both the depth and the breadth of natural languages for two different linguistic tasks: syntactic parsing and joint learning of named entity recognition and coreference resolution.
The syntactic parsing model outperforms current state of the art models by discovering linguistic information shared across languages at the granular level of a sentence. The coreference resolution system is one of the first attempts at joint multilingual modeling of named entity recognition and coreference resolution with limited linguistic resources. It performs second best on three out of four languages when compared to state of the art systems built with rich linguistic resources. Andreea shows that we can simultaneously model both the depth and the breadth of natural languages using the underlying linguistic structure shared across languages.
- published: 13 Nov 2014
- views: 362
Professor Manfred Krifka - Research Seminar Series
- Order: Reorder
- Duration: 55:37
- Updated: 23 Aug 2013
- views: 344
Polarity Particles as Propositional Anaphors
Recently, a number of theories have been proposed to explain polarity particles like yes and no (cf. Kramer & Rawl...
Polarity Particles as Propositional Anaphors
Recently, a number of theories have been proposed to explain polarity particles like yes and no (cf. Kramer & Rawlins 2009, Holmberg 2012, Farkas & Roelofsen 2012). I propose an analysis of polarity particles as anaphoric expressions that pick up propositional discourse referents, and that are themselves of the category of a clause. If the antecedent clauses contain negations, two such discourse referents are introduced, leading to the well-known ambiguities: John isn't coming to the party. -- No, he isn't, but also Yes, he isn't. It will be shown that pragmatic optimization leads to the selection of favourite answers in systems like English (with yes and no) and German (with ja, nein and doch).
wn.com/Professor Manfred Krifka Research Seminar Series
Polarity Particles as Propositional Anaphors
Recently, a number of theories have been proposed to explain polarity particles like yes and no (cf. Kramer & Rawlins 2009, Holmberg 2012, Farkas & Roelofsen 2012). I propose an analysis of polarity particles as anaphoric expressions that pick up propositional discourse referents, and that are themselves of the category of a clause. If the antecedent clauses contain negations, two such discourse referents are introduced, leading to the well-known ambiguities: John isn't coming to the party. -- No, he isn't, but also Yes, he isn't. It will be shown that pragmatic optimization leads to the selection of favourite answers in systems like English (with yes and no) and German (with ja, nein and doch).
- published: 23 Aug 2013
- views: 344
People Recognition - A Leapfrog in Organizing Videos
- Order: Reorder
- Duration: 49:16
- Updated: 18 Oct 2008
- views: 7799
Google Tech Talks
October 2, 2008
ABSTRACT
Shifting from its initial purely entertainment-centric agenda, video is becoming an important media for communicati...
Google Tech Talks
October 2, 2008
ABSTRACT
Shifting from its initial purely entertainment-centric agenda, video is becoming an important media for communication, saving generic facts and personal memories. Viewdle is considering face recognition in real-life videos as one of the three material steps ahead (along with tracking location and timecodes) towards effective organization of videos. The presentation will focus on algorithmic and engineering approaches to address practical challenges of object recognition in low-resolution professional and user-generated videos. One of such approaches is multimodal fusion from several sources of information to achieve best-possible tagging and people-recognition-in-video experience. Viewdle outlines several key principles for "fusing" decisions from combination of heterogeneous sources such as video track, audio track (speech-to-text), OCR data (Optical Character Recognition). A real case of recognition in video taken by smartphone will be demonstrated.
Viewdle presentation builds a bridge from generic problems of information retrieval in video to well-understood products such as video search/discovery, search engine optimization, and organization of mobile videos.
Speaker: Viewdle Representative
Viewdle is a facial-recognition based video indexing and search platform. It combines technology from video, audio, and text indexing to deliver a semantically accurate video search experience. Viewdle is looking inside every frame to identify and deliver factual, contextually-relevant results at a level of granularity not possible with existing video search tools. Viewdle launched in September 2007 at TechCrunch40 in San Francisco where it was one of the 40 presenters selected from 800+ startups Worldwide. Since then the company has expanded beyond white-label B2B offering to consumer-facing widgets and facebook application for organizing personal videos. Viewdle was a "Best Technology Innovation/Achievement" Crunchies 2007 Award finalist. The Company was also selected from among 45 contenders to be one of the eight presenting companies for Venture Night at the LIFT08 conference held in Geneva, Switzerland February 2008. Most recently Viewdle was named the 2008 Start-Up of the Year in Europe by the Plugg Conference. Viewdle technology is currently used by ThomsonReuters.
wn.com/People Recognition A Leapfrog In Organizing Videos
Google Tech Talks
October 2, 2008
ABSTRACT
Shifting from its initial purely entertainment-centric agenda, video is becoming an important media for communication, saving generic facts and personal memories. Viewdle is considering face recognition in real-life videos as one of the three material steps ahead (along with tracking location and timecodes) towards effective organization of videos. The presentation will focus on algorithmic and engineering approaches to address practical challenges of object recognition in low-resolution professional and user-generated videos. One of such approaches is multimodal fusion from several sources of information to achieve best-possible tagging and people-recognition-in-video experience. Viewdle outlines several key principles for "fusing" decisions from combination of heterogeneous sources such as video track, audio track (speech-to-text), OCR data (Optical Character Recognition). A real case of recognition in video taken by smartphone will be demonstrated.
Viewdle presentation builds a bridge from generic problems of information retrieval in video to well-understood products such as video search/discovery, search engine optimization, and organization of mobile videos.
Speaker: Viewdle Representative
Viewdle is a facial-recognition based video indexing and search platform. It combines technology from video, audio, and text indexing to deliver a semantically accurate video search experience. Viewdle is looking inside every frame to identify and deliver factual, contextually-relevant results at a level of granularity not possible with existing video search tools. Viewdle launched in September 2007 at TechCrunch40 in San Francisco where it was one of the 40 presenters selected from 800+ startups Worldwide. Since then the company has expanded beyond white-label B2B offering to consumer-facing widgets and facebook application for organizing personal videos. Viewdle was a "Best Technology Innovation/Achievement" Crunchies 2007 Award finalist. The Company was also selected from among 45 contenders to be one of the eight presenting companies for Venture Night at the LIFT08 conference held in Geneva, Switzerland February 2008. Most recently Viewdle was named the 2008 Start-Up of the Year in Europe by the Plugg Conference. Viewdle technology is currently used by ThomsonReuters.
- published: 18 Oct 2008
- views: 7799
Let's Play Scribblenauts Unlimited [Deutsch] - Part 9 : Der Pandahaltige Metapher-Wald
- Order: Reorder
- Duration: 25:32
- Updated: 16 Feb 2013
- views: 533
Facebook ► http://tiny.cc/8t76ow
Youtube ► http://tiny.cc/tu76ow
- Informationen zum Game / LP -
Spiel/Game : Scribblenauts Unlimited
System/Platform : PC
Ent...
Facebook ► http://tiny.cc/8t76ow
Youtube ► http://tiny.cc/tu76ow
- Informationen zum Game / LP -
Spiel/Game : Scribblenauts Unlimited
System/Platform : PC
Entwickler/Developer : 5th Cell Media
Publisher : Warner Bros. Interactive Entertainment
Sprache/Language : German / Deutsch
Spieler/Player + Kommentar/Commentary : Rivon
Auflösung/Resolution : 1280x720
Genre : Puzzle, Action
Release : 19. November 2012
Ihr habt jetz total lust auf diesen Titel bekommen und wollt ihn unbedingt auch mal zocken? Dann könnt ihr euch das Game hier bei Steam kaufen :
Steam
http://store.steampowered.com/app/218680
- Social Network und son Kram -
Facebook ► http://tiny.cc/8t76ow
Rivon Xfire ► http://tiny.cc/ov76ow
Derfer Xfire ► http://tiny.cc/rw76ow
Rivon Steam ► http://tiny.cc/cx76ow
Derfer Steam ► http://tiny.cc/wx76ow
wn.com/Let's Play Scribblenauts Unlimited Deutsch Part 9 Der Pandahaltige Metapher Wald
Facebook ► http://tiny.cc/8t76ow
Youtube ► http://tiny.cc/tu76ow
- Informationen zum Game / LP -
Spiel/Game : Scribblenauts Unlimited
System/Platform : PC
Entwickler/Developer : 5th Cell Media
Publisher : Warner Bros. Interactive Entertainment
Sprache/Language : German / Deutsch
Spieler/Player + Kommentar/Commentary : Rivon
Auflösung/Resolution : 1280x720
Genre : Puzzle, Action
Release : 19. November 2012
Ihr habt jetz total lust auf diesen Titel bekommen und wollt ihn unbedingt auch mal zocken? Dann könnt ihr euch das Game hier bei Steam kaufen :
Steam
http://store.steampowered.com/app/218680
- Social Network und son Kram -
Facebook ► http://tiny.cc/8t76ow
Rivon Xfire ► http://tiny.cc/ov76ow
Derfer Xfire ► http://tiny.cc/rw76ow
Rivon Steam ► http://tiny.cc/cx76ow
Derfer Steam ► http://tiny.cc/wx76ow
- published: 16 Feb 2013
- views: 533
close fullscreen
- Playlist
- Chat
close fullscreen
- Playlist
- Chat
72:53
EM Works for Pronoun-Anaphora Resolution
Google Tech Talks
October 15, 2008
ABSTRACT
EM (the Expectation Maximization algorithm) ...
published: 17 Oct 2008
EM Works for Pronoun-Anaphora Resolution
EM Works for Pronoun-Anaphora Resolution
- Report rights infringement
- published: 17 Oct 2008
- views: 3752
5:06
AnaphoraResolution1.mp4
MediaHub Start-up and testing of Anaphora Resolution Scenario.
published: 30 Mar 2010
AnaphoraResolution1.mp4
AnaphoraResolution1.mp4
- Report rights infringement
- published: 30 Mar 2010
- views: 140
0:54
Demo of Embodied Anaphor Resolution
This video demonstrates how embodied experience is used to enable an AI to correctly resol...
published: 23 Nov 2009
Demo of Embodied Anaphor Resolution
Demo of Embodied Anaphor Resolution
- Report rights infringement
- published: 23 Nov 2009
- views: 1105
51:00
Mining Unstructured Healthcare Data
Deep Dhillon, Chief Data Scientist at Alliance Health Networks, presents a talk titled "Mi...
published: 16 May 2013
Mining Unstructured Healthcare Data
Mining Unstructured Healthcare Data
- Report rights infringement
- published: 16 May 2013
- views: 2393
15:16
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2 Discussion
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Lingu...
published: 04 Dec 2012
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2 Discussion
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2 Discussion
- Report rights infringement
- published: 04 Dec 2012
- views: 85
42:16
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2
Francesco Mambrini (CHS, USA/DAI, Germany), "Treebanking in the World of Thucydides. Lingu...
published: 04 Dec 2012
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2
Digital Classicist Seminar Berlin (2012/2013) - Seminar 2
- Report rights infringement
- published: 04 Dec 2012
- views: 252
2:33
Anaphora and Epistrophe
A brief explanation of the literary devices anaphora and epistrophe featuring Charles Dick...
published: 13 Apr 2014
Anaphora and Epistrophe
Anaphora and Epistrophe
- Report rights infringement
- published: 13 Apr 2014
- views: 2377
5:34
Anaphora
Dr. Christopher Hanlon and Dr. C.C. Wharram discuss anaphora for the Close Reading Coopera...
published: 03 Jun 2011
Anaphora
Anaphora
- Report rights infringement
- published: 03 Jun 2011
- views: 2755
51:22
Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...
Modeling the Lifespan of Discourse Entities with Application to Coreference Resolution
Dr...
published: 08 Dec 2014
Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...
Dr. Marie-Catherine de Marneffe: Modeling the Lifespan of Discourse Entities with Application to ...
- Report rights infringement
- published: 08 Dec 2014
- views: 143
close fullscreen
- Playlist
- Chat
14:03
Truthseiyer's Minecraft Adventure Ep-015-pt2
Today I nearly finish my trap. The whole drag 'em in and kill 'em works fine, but I seem t...
published: 03 Apr 2011
Truthseiyer's Minecraft Adventure Ep-015-pt2
Truthseiyer's Minecraft Adventure Ep-015-pt2
- Report rights infringement
- published: 03 Apr 2011
- views: 28
close fullscreen
- Playlist
- Chat
27:46
Active Learning for Clinical Coreference Resolution
Tutorial: Tim Miller
Active Learning for Clinical Co-reference Resolution
iDash NLP Annot...
published: 02 Nov 2012
Active Learning for Clinical Coreference Resolution
Active Learning for Clinical Coreference Resolution
- Report rights infringement
- published: 02 Nov 2012
- views: 180
29:41
Coreference Resolution for Morphologically Rich ...
"Coreference Resolution for Morphologically Rich Languages. Adaptation of the Stanford Sys...
published: 22 Dec 2015
Coreference Resolution for Morphologically Rich ...
Coreference Resolution for Morphologically Rich ...
- Report rights infringement
- published: 22 Dec 2015
- views: 41
47:37
Scaling Language Understanding via Joint Multilingual Learning - Andreea Bodnari
See the full post here:
Natural language is a pervasive human skill not yet fully achieva...
published: 13 Nov 2014
Scaling Language Understanding via Joint Multilingual Learning - Andreea Bodnari
Scaling Language Understanding via Joint Multilingual Learning - Andreea Bodnari
- Report rights infringement
- published: 13 Nov 2014
- views: 362
55:37
Professor Manfred Krifka - Research Seminar Series
Polarity Particles as Propositional Anaphors
Recently, a number of theories have been pro...
published: 23 Aug 2013
Professor Manfred Krifka - Research Seminar Series
Professor Manfred Krifka - Research Seminar Series
- Report rights infringement
- published: 23 Aug 2013
- views: 344
49:16
People Recognition - A Leapfrog in Organizing Videos
Google Tech Talks
October 2, 2008
ABSTRACT
Shifting from its initial purely entertainmen...
published: 18 Oct 2008
People Recognition - A Leapfrog in Organizing Videos
People Recognition - A Leapfrog in Organizing Videos
- Report rights infringement
- published: 18 Oct 2008
- views: 7799
25:32
Let's Play Scribblenauts Unlimited [Deutsch] - Part 9 : Der Pandahaltige Metapher-Wald
Facebook ► http://tiny.cc/8t76ow
Youtube ► http://tiny.cc/tu76ow
- Informationen zum Game...
published: 16 Feb 2013
Let's Play Scribblenauts Unlimited [Deutsch] - Part 9 : Der Pandahaltige Metapher-Wald
Let's Play Scribblenauts Unlimited [Deutsch] - Part 9 : Der Pandahaltige Metapher-Wald
- Report rights infringement
- published: 16 Feb 2013
- views: 533
Active Learning for Clinical Coreference Resolutio...
Coreference Resolution for Morphologically Rich .....
Scaling Language Understanding via Joint Multiling...
Professor Manfred Krifka - Research Seminar Series...
People Recognition - A Leapfrog in Organizing Vide...
![Let's Play Scribblenauts Unlimited [Deutsch] - Part 9 : Der](http://web.archive.org./web/20160619174310/http://i.ytimg.com/vi/Q6tknI8EYWw/0.jpg)
Let's Play Scribblenauts Unlimited [Deutsch] - Par...
photo: US National Park Service / K Fink
Woman rescues 5-year-old son from mountain lion's jaws
Edit Business Insider 19 Jun 2016
DENVER (AP) - Summoned by the sound of screams, a Colorado woman raced to her front yard to find a terrifying sight. A mountain lion was hunched over her 5-year-old son, biting him. The woman charged the animal, yanked away one of its paws and discovered her son's whole head was in its mouth. She didn't back down. "She was able to pry the cat's jaws open," Pitkin County Sheriff's Deputy Michael Buglione said. "She's a hero." ... ....
photo: WN / Ahmed Deeb
Israel: Water as a tool to dominate Palestinians
Edit Al Jazeera 19 Jun 2016
As temperatures rise and summer months approach, yet again this year, thousands of Palestinians in the occupied West Bank are being deprived of their most basic need - access to water - as the Israeli national water company Mekorot restricted the water supply to villages and towns in northern West Bank ... As a result, thousands of Palestinians are unable to access sufficient water supplies and became water-dependent on Israel ... READ MORE....
photo: AP / Frank Augstein
Jo Cox death has taken 'momentum' away from Brexit, says Farage
Edit The Independent 19 Jun 2016
The death of Labour MP Jo Cox has taken some of the momentum out of the campaign to get Britain out of the European Union, Ukip leader Nigel Farage has suggested ... Leading Leave campaigner Michael Gove also distanced himself from the image, saying he "shuddered" when he saw it ... "I think we have momentum," said one of the leading advocates of Brexit. "We did have momentum until this terrible tragedy ... What we saw was an act of terrorism."....
photo: AP / Matt Slocum
Copa America: Eduardo Vargas scores four goals, defending champions Chile rout Mexico
Edit Indian Express 19 Jun 2016
Chile registered a massive 7-0 win over Mexico. (Source. USA Today Sports). Eduardo Vargas leapt in the air and pumped his right fist in triumph as his shaggy hair swayed in the Bay Area breeze, only further demoralizing Mexico on the worst night in El Tri’s storied soccer history ... “First, I’d like to offer my most heartfelt apologies to the Mexican fans,” Mexico coach Juan Carlos Osorio said ... ....
photo: AP / Frank Augstein
'Angry' Julian Assange starts fifth year living in Ecuador’s London embassy
Edit The Guardian 19 Jun 2016
Noam Chomsky, Yanis Varoufakis, Ai Weiwei, Michael Moore and Ken Loach are among prominent voices demanding WikiLeaks founder be allowed to walk free. WikiLeaks founder Julian Assange starts his fifth year camped out in the Ecuadoran embassy in London on Sunday, an occasion his supporters said they would mark with events celebrating whistleblowers ... Related. Swedish court upholds Julian Assange arrest warrant ... “We live in a critical time....
« back to news headlines
To be or not to be with Shiv Sena, BJP state unit to decide today
Edit DNA India 19 Jun 2016
The two-day executive of the BJP state unit, which would conclude on Sunday, is likely to pass the political resolution, expressing the need to strengthen the party in such a way that it can get rid of Shiv Sena as early as possible ... The same is likely to reflect in the political resolution ... Resolution delayed. The political resolution was to be taken ......
AIADMK demands funds from Centre
Edit Deccan Chronicle 19 Jun 2016
The AIADMK executive on Saturday discussed strategies to be taken up for the ensuing local body polls and passed resolutions thanking Tamil Nadu voters, particularly those residing in Chief Minister J ... were the key issues discussed and resolutions adopted....
Hint to use Isro data for monitoring
Edit The Times of India 19 Jun 2016
The Pune Municipal Corporation is consulting Space Application Centre, Indian Space Research Organisation, Ahmedabad, for using high resolution remote sensing picture data to keep a tab on encroachments and creating a 3D model of the city for estimating rooftop energy potential....
Jubilant AIADMK fixes win in civic polls as next target
Edit Indian Express 19 Jun 2016
“This executive committee takes a vow to bag a massive victory in the coming local body elections,” a resolution adopted in the meet said ... A resolution lauded Jayalalithaa for becoming Chief Minister for the sixth time and thanked her and the people for ensuring the party’s victory in the May 16 Assembly election ... Other resolutions ......
EU Bank Resolution Plan Ignored in Denmark After Bail-In Trauma
Edit Bloomberg 19 Jun 2016
Denmark’s financial regulator says it won’t abide by a key principle in Europe’s Bank Recovery and Resolution Directive as the nation carves out its own vision for how to tackle insolvencies ... It was the first country in Europe to adopt a resolution framework in 2010 ... “We realized that we had no resolution framework -- nobody did -- apart from an ordinary bankruptcy.” ... Creditors shouldn’t lose more if resolution is chosen.MREL Questions....
BJP asks Mamata to put an end to attacks on opposition
Edit Deccan Chronicle 19 Jun 2016
Urging West Bengal Chief Minister Mamata Banerjee to put an end to such attacks on opposition, the meeting on Saturday passed a resolution which called for a movement to stop such atrocities ... The resolution, however, appreciated the steps already taken by the Awami League government in this regard....
Proposed charter revision would change how city disposes of property
Edit Springfield News-Sun 19 Jun 2016
The proposed changes are planned for the November ballot ... Under the new amendment, routine business resolutions with no funding attached would get only one reading. Funded resolutions and ordinances would require at least two readings, while permanent resolutions would still require three readings ... Currently, a separate advertisement for every adopted city resolution or ordinance is required to be published in the newspaper ... ....
BJP to launch movement in West Bengal to protest against post-poll violence
Edit Indian Express 19 Jun 2016
Urging West Bengal Chief Minister Mamata Banerjee to put an end to such attacks on opposition, the meeting on June 18 passed a resolution which called for a movement to stop such atrocities. In another resolution, the meeting referred to the attack on minorities in neighbouring Bangladesh and called upon the Sheikh Hasina government to put a halt to such attacks....
When 97 were killed in Iraq Alone
Edit Dissident Voice 19 Jun 2016
Twas the twelfth. of June1. and all through. the house ... was stirring ... Excited, ... Then war was escalated on 14 September 2001 with the adoption of the Authorization of the Use of Military Force Against Terrorists (AUMF), a circumvention not unlike the 10 August 1964 joint resolution “to promote the maintenance of international peace and security in southeast Asis”, aka the “Tonkin Gulf Resolution”. ....
Governance in focus: Be good, because investors are watching more closely
Edit The National 19 Jun 2016
Despite that, corporate governance codes, such as those outlined in an updated UAE resolution issued in late April, have certain themes in common ... Most companies listed on the Abu Dhabi Securities Exchange and Dubai Financial Market were from late April 2010 obliged to comply with a 2009 resolution, regulated by the UAE Securities and Commodities Authority (SCA), covering governance rules and corporate discipline standards....
Outcome of the 26th Plenary Meeting of the Nuclear Suppliers G... (Ministry of Foreign Affairs and Trade of the Republic of Korea)
Edit Public Technologies 19 Jun 2016
(Source. Ministry of Foreign Affairs and Trade of the Republic of Korea). 1 ... 2 ... 3 ... ◦ The Minister will highlight the importance of the thorough implementation of UN Security Council Resolution 2270 for the resolution of the North Korean nuclear issue and will ask the NSG member states to double their efforts to further rein in the supply chain for the North Korean nuclear program. 4 ... * unofficial translation ... (noodl. 34165747) ....
East Bay steelmaker USS-POSCO in international trade dispute bind
Edit Topix 19 Jun 2016
The wrong resolution, USS-POSCO officials and supporters contend, could disrupt the Pittsburg plant's supply of Korean "hot-rolled" steel, possibly jeopardizing some or all of the plant's 620 jobs and causing ripple effects through East County's industrial sector ... ....
Cabinet to allocate more funds to haredi schools, settlements
Edit The Jerusalem Post 19 Jun 2016
According to the explanation in the resolution, the worsening security situation had an influence on life in the settlements in a number of areas ... ....
- 1
- 2
- 3
- 4
- 5
- Next page »
