�ݺ�ߣshows by User: ascherp

�ݺ�ߣshows by User: ascherp / http://www.slideshare.net/images/logo.gif �ݺ�ߣshows by User: ascherp / Mon, 24 Jan 2022 21:20:00 GMT �ݺ�ߣShare feed for �ݺ�ߣshows by User: ascherp Analysis of GraphSum's Attention Weights to Improve the Explainability of Multi-Document Summarization /slideshow/analysis-of-graphsums-attention-weights-to-improve-the-explainability-of-multidocument-summarization/251047624 graphsumpresentation02final-220124212001
�ݺ�ߣs of our presentation @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4]]>
�ݺ�ߣs of our presentation @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4]]> Mon, 24 Jan 2022 21:20:00 GMT /slideshow/analysis-of-graphsums-attention-weights-to-improve-the-explainability-of-multidocument-summarization/251047624 ascherp@slideshare.net(ascherp) Analysis of GraphSum's Attention Weights to Improve the Explainability of Multi-Document Summarization ascherp �ݺ�ߣs of our presentation @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/graphsumpresentation02final-220124212001-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs of our presentation @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4

Analysis of GraphSum's Attention Weights to Improve the Explainability of Multi-Document Summarization from Ansgar Scherp

]]> 75 0 https://cdn.slidesharecdn.com/ss_thumbnails/graphsumpresentation02final-220124212001-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 STEREO: A Pipeline for Extracting Experiment Statistics, Conditions, and Topics from Scientific Papers /slideshow/stereo-a-pipeline-for-extracting-experiment-statistics-conditions-and-topics-from-scientific-papers/251047618 stereoiiwas-presentation-220124211535
Presentation for our paper @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4]]>
Presentation for our paper @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4]]> Mon, 24 Jan 2022 21:15:35 GMT /slideshow/stereo-a-pipeline-for-extracting-experiment-statistics-conditions-and-topics-from-scientific-papers/251047618 ascherp@slideshare.net(ascherp) STEREO: A Pipeline for Extracting Experiment Statistics, Conditions, and Topics from Scientific Papers ascherp Presentation for our paper @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/stereoiiwas-presentation-220124211535-thumbnail.jpg?width=120&height=120&fit=bounds" /> Presentation for our paper @iiWAS2021: The 23rd International Conference on Information Integration and Web Intelligence, Linz, Austria, 29 November 2021 - 1 December 2021. ACM 2021, ISBN 978-1-4503-9556-4

STEREO: A Pipeline for Extracting Experiment Statistics, Conditions, and Topics from Scientific Papers from Ansgar Scherp

]]> 44 0 https://cdn.slidesharecdn.com/ss_thumbnails/stereoiiwas-presentation-220124211535-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Text Localization in Scientific Figures using Fully Convolutional Neural Networks on Limited Training Data /slideshow/text-localization-in-scientific-figures-using-fully-convolutional-neural-networks-on-limited-training-data/176331413 jessenboeschenscherp-nntextextraction-190926130309
Text extraction from scientific figures has been addressed in the past by different unsupervised approaches due to the limited amount of training data. Motivated by the recent advances in Deep Learning, we propose a two-step neural-network-based pipeline to localize and extract text using Fully Convolutional Networks. We improve the localization of the text bounding boxes by applying a novel combination of a Residual Network with the Region Proposal Network based on Faster R-CNN. The predicted bounding boxes are further pre-processed and used as input to the of-the-shelf optical character recognition engine Tesseract 4.0. We evaluate our improved text localization method on five different datasets of scientific figures and compare it with the best unsupervised pipeline. Since only limited training data is available, we further experiment with different data augmentation techniques for increasing the size of the training datasets and demonstrate their positive impact. We use Average Precision and F1 measure to assess the text localization results. In addition, we apply Gestalt Pattern Matching and Levenshtein Distance for evaluating the quality of the recognized text. Our extensive experiments show that our new pipeline based on neural networks outperforms the best unsupervised approach by a large margin of 19-20%. ]]>
Text extraction from scientific figures has been addressed in the past by different unsupervised approaches due to the limited amount of training data. Motivated by the recent advances in Deep Learning, we propose a two-step neural-network-based pipeline to localize and extract text using Fully Convolutional Networks. We improve the localization of the text bounding boxes by applying a novel combination of a Residual Network with the Region Proposal Network based on Faster R-CNN. The predicted bounding boxes are further pre-processed and used as input to the of-the-shelf optical character recognition engine Tesseract 4.0. We evaluate our improved text localization method on five different datasets of scientific figures and compare it with the best unsupervised pipeline. Since only limited training data is available, we further experiment with different data augmentation techniques for increasing the size of the training datasets and demonstrate their positive impact. We use Average Precision and F1 measure to assess the text localization results. In addition, we apply Gestalt Pattern Matching and Levenshtein Distance for evaluating the quality of the recognized text. Our extensive experiments show that our new pipeline based on neural networks outperforms the best unsupervised approach by a large margin of 19-20%. ]]> Thu, 26 Sep 2019 13:03:09 GMT /slideshow/text-localization-in-scientific-figures-using-fully-convolutional-neural-networks-on-limited-training-data/176331413 ascherp@slideshare.net(ascherp) Text Localization in Scientific Figures using Fully Convolutional Neural Networks on Limited Training Data ascherp Text extraction from scientific figures has been addressed in the past by different unsupervised approaches due to the limited amount of training data. Motivated by the recent advances in Deep Learning, we propose a two-step neural-network-based pipeline to localize and extract text using Fully Convolutional Networks. We improve the localization of the text bounding boxes by applying a novel combination of a Residual Network with the Region Proposal Network based on Faster R-CNN. The predicted bounding boxes are further pre-processed and used as input to the of-the-shelf optical character recognition engine Tesseract 4.0. We evaluate our improved text localization method on five different datasets of scientific figures and compare it with the best unsupervised pipeline. Since only limited training data is available, we further experiment with different data augmentation techniques for increasing the size of the training datasets and demonstrate their positive impact. We use Average Precision and F1 measure to assess the text localization results. In addition, we apply Gestalt Pattern Matching and Levenshtein Distance for evaluating the quality of the recognized text. Our extensive experiments show that our new pipeline based on neural networks outperforms the best unsupervised approach by a large margin of 19-20%. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/jessenboeschenscherp-nntextextraction-190926130309-thumbnail.jpg?width=120&height=120&fit=bounds" /> Text extraction from scientific figures has been addressed in the past by different unsupervised approaches due to the limited amount of training data. Motivated by the recent advances in Deep Learning, we propose a two-step neural-network-based pipeline to localize and extract text using Fully Convolutional Networks. We improve the localization of the text bounding boxes by applying a novel combination of a Residual Network with the Region Proposal Network based on Faster R-CNN. The predicted bounding boxes are further pre-processed and used as input to the of-the-shelf optical character recognition engine Tesseract 4.0. We evaluate our improved text localization method on five different datasets of scientific figures and compare it with the best unsupervised pipeline. Since only limited training data is available, we further experiment with different data augmentation techniques for increasing the size of the training datasets and demonstrate their positive impact. We use Average Precision and F1 measure to assess the text localization results. In addition, we apply Gestalt Pattern Matching and Levenshtein Distance for evaluating the quality of the recognized text. Our extensive experiments show that our new pipeline based on neural networks outperforms the best unsupervised approach by a large margin of 19-20%.

Text Localization in Scientific Figures using Fully Convolutional Neural Networks on Limited Training Data from Ansgar Scherp

]]> 318 0 https://cdn.slidesharecdn.com/ss_thumbnails/jessenboeschenscherp-nntextextraction-190926130309-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 A Comparison of Approaches for Automated Text Extraction from Scholarly Figures /slideshow/a-comparison-of-approaches-for-automated-text-extraction-from-scholarly-figures/87611286 mmmpresentationclean-180209094345
So far, there has not been a comparative evaluation of different approaches for text extraction from scholarly figures. In order to fill this gap, we have defined a generic pipeline for text extraction that abstracts from the existing approaches as documented in the literature. In this paper, we use this generic pipeline to systematically evaluate and compare 32 configurations for text extraction over four datasets of scholarly figures of different origin and characteristics. In total, our experiments have been run over more than 400 manually labeled figures. The experimental results show that the approach BS-4OS results in the best F-measure of 0.67 for the Text Location Detection and the best average Levenshtein Distance of 4.71 between the recognized text and the gold standard on all four datasets using the Ocropy OCR engine.]]>
So far, there has not been a comparative evaluation of different approaches for text extraction from scholarly figures. In order to fill this gap, we have defined a generic pipeline for text extraction that abstracts from the existing approaches as documented in the literature. In this paper, we use this generic pipeline to systematically evaluate and compare 32 configurations for text extraction over four datasets of scholarly figures of different origin and characteristics. In total, our experiments have been run over more than 400 manually labeled figures. The experimental results show that the approach BS-4OS results in the best F-measure of 0.67 for the Text Location Detection and the best average Levenshtein Distance of 4.71 between the recognized text and the gold standard on all four datasets using the Ocropy OCR engine.]]> Fri, 09 Feb 2018 09:43:44 GMT /slideshow/a-comparison-of-approaches-for-automated-text-extraction-from-scholarly-figures/87611286 ascherp@slideshare.net(ascherp) A Comparison of Approaches for Automated Text Extraction from Scholarly Figures ascherp So far, there has not been a comparative evaluation of different approaches for text extraction from scholarly figures. In order to fill this gap, we have defined a generic pipeline for text extraction that abstracts from the existing approaches as documented in the literature. In this paper, we use this generic pipeline to systematically evaluate and compare 32 configurations for text extraction over four datasets of scholarly figures of different origin and characteristics. In total, our experiments have been run over more than 400 manually labeled figures. The experimental results show that the approach BS-4OS results in the best F-measure of 0.67 for the Text Location Detection and the best average Levenshtein Distance of 4.71 between the recognized text and the gold standard on all four datasets using the Ocropy OCR engine. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/mmmpresentationclean-180209094345-thumbnail.jpg?width=120&height=120&fit=bounds" /> So far, there has not been a comparative evaluation of different approaches for text extraction from scholarly figures. In order to fill this gap, we have defined a generic pipeline for text extraction that abstracts from the existing approaches as documented in the literature. In this paper, we use this generic pipeline to systematically evaluate and compare 32 configurations for text extraction over four datasets of scholarly figures of different origin and characteristics. In total, our experiments have been run over more than 400 manually labeled figures. The experimental results show that the approach BS-4OS results in the best F-measure of 0.67 for the Text Location Detection and the best average Levenshtein Distance of 4.71 between the recognized text and the gold standard on all four datasets using the Ocropy OCR engine.

A Comparison of Approaches for Automated Text Extraction from Scholarly Figures from Ansgar Scherp

]]> 601 1 https://cdn.slidesharecdn.com/ss_thumbnails/mmmpresentationclean-180209094345-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 About Multimedia Presentation Generation and Multimedia Metadata: From Synthesis to Analysis, and Back? /slideshow/about-multimedia-presentation-generation-and-multimedia-metadata-from-synthesis-to-analysis-and-back/67515173 talk-acm-mm-final-161021203143
ACM SIGMM Rising Stars Symposium The ACM SIGMM Rising Stars Symposium, inaugurated in 2015, will highlight plenary presentations of six selected rising SIGMM members on their vision and research achievements, and dialogs with senior members about the future of multimedia research. See: http://www.acmmm.org/2016/?page_id=706]]>
ACM SIGMM Rising Stars Symposium The ACM SIGMM Rising Stars Symposium, inaugurated in 2015, will highlight plenary presentations of six selected rising SIGMM members on their vision and research achievements, and dialogs with senior members about the future of multimedia research. See: http://www.acmmm.org/2016/?page_id=706]]> Fri, 21 Oct 2016 20:31:43 GMT /slideshow/about-multimedia-presentation-generation-and-multimedia-metadata-from-synthesis-to-analysis-and-back/67515173 ascherp@slideshare.net(ascherp) About Multimedia Presentation Generation and Multimedia Metadata: From Synthesis to Analysis, and Back? ascherp ACM SIGMM Rising Stars Symposium The ACM SIGMM Rising Stars Symposium, inaugurated in 2015, will highlight plenary presentations of six selected rising SIGMM members on their vision and research achievements, and dialogs with senior members about the future of multimedia research. See: http://www.acmmm.org/2016/?page_id=706 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/talk-acm-mm-final-161021203143-thumbnail.jpg?width=120&height=120&fit=bounds" /> ACM SIGMM Rising Stars Symposium The ACM SIGMM Rising Stars Symposium, inaugurated in 2015, will highlight plenary presentations of six selected rising SIGMM members on their vision and research achievements, and dialogs with senior members about the future of multimedia research. See: http://www.acmmm.org/2016/?page_id=706

About Multimedia Presentation Generation and Multimedia Metadata: From Synthesis to Analysis, and Back? from Ansgar Scherp

]]> 1245 6 https://cdn.slidesharecdn.com/ss_thumbnails/talk-acm-mm-final-161021203143-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Mining and Managing Large-scale Linked Open Data /slideshow/mining-and-managing-largescale-linked-open-data/62377363 gvdbtalkfinal-160525090150
Linked Open Data (LOD) is about publishing and interlinking data of different origin and purpose on the web. The Resource Description Framework (RDF) is used to describe data on the LOD cloud. In contrast to relational databases, RDF does not provide a fixed, pre-defined schema. Rather, RDF allows for flexibly modeling the data schema by attaching RDF types and properties to the entities. Our schema-level index called SchemEX allows for searching in large-scale RDF graph data. The index can be efficiently computed with reasonable accuracy over large-scale data sets with billions of RDF triples, the smallest information unit on the LOD cloud. SchemEX is highly needed as the size of the LOD cloud quickly increases. Due to the evolution of the LOD cloud, one observes frequent changes of the data. We show that also the data schema changes in terms of combinations of RDF types and properties. As changes cannot capture the dynamics of the LOD cloud, current work includes temporal clustering and finding periodicities in entity dynamics over large-scale snapshots of the LOD cloud with about 100 million triples per week for more than three years.]]>
Linked Open Data (LOD) is about publishing and interlinking data of different origin and purpose on the web. The Resource Description Framework (RDF) is used to describe data on the LOD cloud. In contrast to relational databases, RDF does not provide a fixed, pre-defined schema. Rather, RDF allows for flexibly modeling the data schema by attaching RDF types and properties to the entities. Our schema-level index called SchemEX allows for searching in large-scale RDF graph data. The index can be efficiently computed with reasonable accuracy over large-scale data sets with billions of RDF triples, the smallest information unit on the LOD cloud. SchemEX is highly needed as the size of the LOD cloud quickly increases. Due to the evolution of the LOD cloud, one observes frequent changes of the data. We show that also the data schema changes in terms of combinations of RDF types and properties. As changes cannot capture the dynamics of the LOD cloud, current work includes temporal clustering and finding periodicities in entity dynamics over large-scale snapshots of the LOD cloud with about 100 million triples per week for more than three years.]]> Wed, 25 May 2016 09:01:50 GMT /slideshow/mining-and-managing-largescale-linked-open-data/62377363 ascherp@slideshare.net(ascherp) Mining and Managing Large-scale Linked Open Data ascherp Linked Open Data (LOD) is about publishing and interlinking data of different origin and purpose on the web. The Resource Description Framework (RDF) is used to describe data on the LOD cloud. In contrast to relational databases, RDF does not provide a fixed, pre-defined schema. Rather, RDF allows for flexibly modeling the data schema by attaching RDF types and properties to the entities. Our schema-level index called SchemEX allows for searching in large-scale RDF graph data. The index can be efficiently computed with reasonable accuracy over large-scale data sets with billions of RDF triples, the smallest information unit on the LOD cloud. SchemEX is highly needed as the size of the LOD cloud quickly increases. Due to the evolution of the LOD cloud, one observes frequent changes of the data. We show that also the data schema changes in terms of combinations of RDF types and properties. As changes cannot capture the dynamics of the LOD cloud, current work includes temporal clustering and finding periodicities in entity dynamics over large-scale snapshots of the LOD cloud with about 100 million triples per week for more than three years. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/gvdbtalkfinal-160525090150-thumbnail.jpg?width=120&height=120&fit=bounds" /> Linked Open Data (LOD) is about publishing and interlinking data of different origin and purpose on the web. The Resource Description Framework (RDF) is used to describe data on the LOD cloud. In contrast to relational databases, RDF does not provide a fixed, pre-defined schema. Rather, RDF allows for flexibly modeling the data schema by attaching RDF types and properties to the entities. Our schema-level index called SchemEX allows for searching in large-scale RDF graph data. The index can be efficiently computed with reasonable accuracy over large-scale data sets with billions of RDF triples, the smallest information unit on the LOD cloud. SchemEX is highly needed as the size of the LOD cloud quickly increases. Due to the evolution of the LOD cloud, one observes frequent changes of the data. We show that also the data schema changes in terms of combinations of RDF types and properties. As changes cannot capture the dynamics of the LOD cloud, current work includes temporal clustering and finding periodicities in entity dynamics over large-scale snapshots of the LOD cloud with about 100 million triples per week for more than three years.

Mining and Managing Large-scale Linked Open Data from Ansgar Scherp

]]> 1408 4 https://cdn.slidesharecdn.com/ss_thumbnails/gvdbtalkfinal-160525090150-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Knowledge Discovery in Social Media and Scientific Digital Libraries /slideshow/knowledge-discovery-in-social-media-and-scientific-digital-libraries/58054556 darmstadt-talk-160209142503
The talk presents selected results of our research in the area of text and data mining in social media and scientific literature. (1) First, we consider the area of classifying microblogging postings like tweets on Twitter. Typically, the classification results are evaluated against a gold standard, which is either the hashtags of the tweets’ authors or manual annotations. We claim that there are fundamental differences between these two kinds of gold standard classifications and conducted an experiment with 163 participants to manually classify tweets from ten topics. Our results show that the human annotators are more likely to classify tweets like other human annotators than like the tweets’ authors (i. e., the hashtags). This may influence the evaluation of classification methods like LDA and we argue that researchers should reflect the kind of gold standard used when interpreting their results. (2) Second, we present a framework for semantic document annotation that aims to compare different existing as well as new annotation strategies. For entity detection, we compare semantic taxonomies, trigrams, RAKE, and LDA. For concept activation, we cover a set of statistical, hierarchy-based, and graph-based methods. The strategies are evaluated over 100,000 manually labeled scientific documents from economics, politics, and computer science. (3) Finally, we present a processing pipeline for extracting text of varying size, rotation, color, and emphases from scholarly figures. The pipeline does not need training nor does it make any assumptions about the characteristics of the scholarly figures. We conducted a preliminary evaluation with 121 figures from a broad range of illustration types. URL: https://www.ukp.tu-darmstadt.de/ukp-home/news-singleview/artikel/guest-speaker-ansgar-scherp/]]>
The talk presents selected results of our research in the area of text and data mining in social media and scientific literature. (1) First, we consider the area of classifying microblogging postings like tweets on Twitter. Typically, the classification results are evaluated against a gold standard, which is either the hashtags of the tweets’ authors or manual annotations. We claim that there are fundamental differences between these two kinds of gold standard classifications and conducted an experiment with 163 participants to manually classify tweets from ten topics. Our results show that the human annotators are more likely to classify tweets like other human annotators than like the tweets’ authors (i. e., the hashtags). This may influence the evaluation of classification methods like LDA and we argue that researchers should reflect the kind of gold standard used when interpreting their results. (2) Second, we present a framework for semantic document annotation that aims to compare different existing as well as new annotation strategies. For entity detection, we compare semantic taxonomies, trigrams, RAKE, and LDA. For concept activation, we cover a set of statistical, hierarchy-based, and graph-based methods. The strategies are evaluated over 100,000 manually labeled scientific documents from economics, politics, and computer science. (3) Finally, we present a processing pipeline for extracting text of varying size, rotation, color, and emphases from scholarly figures. The pipeline does not need training nor does it make any assumptions about the characteristics of the scholarly figures. We conducted a preliminary evaluation with 121 figures from a broad range of illustration types. URL: https://www.ukp.tu-darmstadt.de/ukp-home/news-singleview/artikel/guest-speaker-ansgar-scherp/]]> Tue, 09 Feb 2016 14:25:03 GMT /slideshow/knowledge-discovery-in-social-media-and-scientific-digital-libraries/58054556 ascherp@slideshare.net(ascherp) Knowledge Discovery in Social Media and Scientific Digital Libraries ascherp The talk presents selected results of our research in the area of text and data mining in social media and scientific literature. (1) First, we consider the area of classifying microblogging postings like tweets on Twitter. Typically, the classification results are evaluated against a gold standard, which is either the hashtags of the tweets’ authors or manual annotations. We claim that there are fundamental differences between these two kinds of gold standard classifications and conducted an experiment with 163 participants to manually classify tweets from ten topics. Our results show that the human annotators are more likely to classify tweets like other human annotators than like the tweets’ authors (i. e., the hashtags). This may influence the evaluation of classification methods like LDA and we argue that researchers should reflect the kind of gold standard used when interpreting their results. (2) Second, we present a framework for semantic document annotation that aims to compare different existing as well as new annotation strategies. For entity detection, we compare semantic taxonomies, trigrams, RAKE, and LDA. For concept activation, we cover a set of statistical, hierarchy-based, and graph-based methods. The strategies are evaluated over 100,000 manually labeled scientific documents from economics, politics, and computer science. (3) Finally, we present a processing pipeline for extracting text of varying size, rotation, color, and emphases from scholarly figures. The pipeline does not need training nor does it make any assumptions about the characteristics of the scholarly figures. We conducted a preliminary evaluation with 121 figures from a broad range of illustration types. URL: https://www.ukp.tu-darmstadt.de/ukp-home/news-singleview/artikel/guest-speaker-ansgar-scherp/ <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/darmstadt-talk-160209142503-thumbnail.jpg?width=120&height=120&fit=bounds" /> The talk presents selected results of our research in the area of text and data mining in social media and scientific literature. (1) First, we consider the area of classifying microblogging postings like tweets on Twitter. Typically, the classification results are evaluated against a gold standard, which is either the hashtags of the tweets’ authors or manual annotations. We claim that there are fundamental differences between these two kinds of gold standard classifications and conducted an experiment with 163 participants to manually classify tweets from ten topics. Our results show that the human annotators are more likely to classify tweets like other human annotators than like the tweets’ authors (i. e., the hashtags). This may influence the evaluation of classification methods like LDA and we argue that researchers should reflect the kind of gold standard used when interpreting their results. (2) Second, we present a framework for semantic document annotation that aims to compare different existing as well as new annotation strategies. For entity detection, we compare semantic taxonomies, trigrams, RAKE, and LDA. For concept activation, we cover a set of statistical, hierarchy-based, and graph-based methods. The strategies are evaluated over 100,000 manually labeled scientific documents from economics, politics, and computer science. (3) Finally, we present a processing pipeline for extracting text of varying size, rotation, color, and emphases from scholarly figures. The pipeline does not need training nor does it make any assumptions about the characteristics of the scholarly figures. We conducted a preliminary evaluation with 121 figures from a broad range of illustration types. URL: https://www.ukp.tu-darmstadt.de/ukp-home/news-singleview/artikel/guest-speaker-ansgar-scherp/

Knowledge Discovery in Social Media and Scientific Digital Libraries from Ansgar Scherp

]]> 1538 7 https://cdn.slidesharecdn.com/ss_thumbnails/darmstadt-talk-160209142503-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 A Comparison of Different Strategies for Automated Semantic Document Annotation /slideshow/a-comparison-of-different-strategies-for-automated-semantic-document-annotation/53980960 2015-kcap-slides-151015153349-lva1-app6891
We introduce a framework for automated semantic document annotation that is composed of four processes, namely concept extraction, concept activation, annotation selection, and evaluation. The framework is used to implement and compare different annotation strategies motivated by the literature. For concept extraction, we apply entity detection with semantic hierarchical knowledge bases, Tri-gram, RAKE, and LDA. For concept activation, we compare a set of statistical, hierarchy-based, and graph-based methods. For selecting annotations, we compare top-k as well as kNN. In total, we define 43 different strategies including novel combinations like using graph-based activation with kNN. We have evaluated the strategies using three different datasets of varying size from three scientific disciplines (economics, politics, and computer science) that contain 100, 000 manually labeled documents in total. We obtain the best results on all three datasets by our novel combination of entity detection with graph-based activation (e.g., HITS and Degree) and kNN. For the economic and political science datasets, the best F-measure is .39 and .28, respectively. For the computer science dataset, the maximum F-measure of .33 can be reached. The experiments are the by far largest on scholarly content annotation, which typically are up to a few hundred documents per dataset only. Gregor Große-Bölting, Chifumi Nishioka, and Ansgar Scherp. 2015. A Comparison of Different Strategies for Automated Semantic Document Annotation. In Proceedings of the 8th International Conference on Knowledge Capture (K-CAP 2015). ACM, New York, NY, USA, , Article 8 , 8 pages. DOI=http://dx.doi.org/10.1145/2815833.2815838 ]]>
We introduce a framework for automated semantic document annotation that is composed of four processes, namely concept extraction, concept activation, annotation selection, and evaluation. The framework is used to implement and compare different annotation strategies motivated by the literature. For concept extraction, we apply entity detection with semantic hierarchical knowledge bases, Tri-gram, RAKE, and LDA. For concept activation, we compare a set of statistical, hierarchy-based, and graph-based methods. For selecting annotations, we compare top-k as well as kNN. In total, we define 43 different strategies including novel combinations like using graph-based activation with kNN. We have evaluated the strategies using three different datasets of varying size from three scientific disciplines (economics, politics, and computer science) that contain 100, 000 manually labeled documents in total. We obtain the best results on all three datasets by our novel combination of entity detection with graph-based activation (e.g., HITS and Degree) and kNN. For the economic and political science datasets, the best F-measure is .39 and .28, respectively. For the computer science dataset, the maximum F-measure of .33 can be reached. The experiments are the by far largest on scholarly content annotation, which typically are up to a few hundred documents per dataset only. Gregor Große-Bölting, Chifumi Nishioka, and Ansgar Scherp. 2015. A Comparison of Different Strategies for Automated Semantic Document Annotation. In Proceedings of the 8th International Conference on Knowledge Capture (K-CAP 2015). ACM, New York, NY, USA, , Article 8 , 8 pages. DOI=http://dx.doi.org/10.1145/2815833.2815838 ]]> Thu, 15 Oct 2015 15:33:49 GMT /slideshow/a-comparison-of-different-strategies-for-automated-semantic-document-annotation/53980960 ascherp@slideshare.net(ascherp) A Comparison of Different Strategies for Automated Semantic Document Annotation ascherp We introduce a framework for automated semantic document annotation that is composed of four processes, namely concept extraction, concept activation, annotation selection, and evaluation. The framework is used to implement and compare different annotation strategies motivated by the literature. For concept extraction, we apply entity detection with semantic hierarchical knowledge bases, Tri-gram, RAKE, and LDA. For concept activation, we compare a set of statistical, hierarchy-based, and graph-based methods. For selecting annotations, we compare top-k as well as kNN. In total, we define 43 different strategies including novel combinations like using graph-based activation with kNN. We have evaluated the strategies using three different datasets of varying size from three scientific disciplines (economics, politics, and computer science) that contain 100, 000 manually labeled documents in total. We obtain the best results on all three datasets by our novel combination of entity detection with graph-based activation (e.g., HITS and Degree) and kNN. For the economic and political science datasets, the best F-measure is .39 and .28, respectively. For the computer science dataset, the maximum F-measure of .33 can be reached. The experiments are the by far largest on scholarly content annotation, which typically are up to a few hundred documents per dataset only. Gregor Große-Bölting, Chifumi Nishioka, and Ansgar Scherp. 2015. A Comparison of Different Strategies for Automated Semantic Document Annotation. In Proceedings of the 8th International Conference on Knowledge Capture (K-CAP 2015). ACM, New York, NY, USA, , Article 8 , 8 pages. DOI=http://dx.doi.org/10.1145/2815833.2815838 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/2015-kcap-slides-151015153349-lva1-app6891-thumbnail.jpg?width=120&height=120&fit=bounds" /> We introduce a framework for automated semantic document annotation that is composed of four processes, namely concept extraction, concept activation, annotation selection, and evaluation. The framework is used to implement and compare different annotation strategies motivated by the literature. For concept extraction, we apply entity detection with semantic hierarchical knowledge bases, Tri-gram, RAKE, and LDA. For concept activation, we compare a set of statistical, hierarchy-based, and graph-based methods. For selecting annotations, we compare top-k as well as kNN. In total, we define 43 different strategies including novel combinations like using graph-based activation with kNN. We have evaluated the strategies using three different datasets of varying size from three scientific disciplines (economics, politics, and computer science) that contain 100, 000 manually labeled documents in total. We obtain the best results on all three datasets by our novel combination of entity detection with graph-based activation (e.g., HITS and Degree) and kNN. For the economic and political science datasets, the best F-measure is .39 and .28, respectively. For the computer science dataset, the maximum F-measure of .33 can be reached. The experiments are the by far largest on scholarly content annotation, which typically are up to a few hundred documents per dataset only. Gregor Große-Bölting, Chifumi Nishioka, and Ansgar Scherp. 2015. A Comparison of Different Strategies for Automated Semantic Document Annotation. In Proceedings of the 8th International Conference on Knowledge Capture (K-CAP 2015). ACM, New York, NY, USA, , Article 8 , 8 pages. DOI=http://dx.doi.org/10.1145/2815833.2815838

A Comparison of Different Strategies for Automated Semantic Document Annotation from Ansgar Scherp

]]> 1014 5 https://cdn.slidesharecdn.com/ss_thumbnails/2015-kcap-slides-151015153349-lva1-app6891-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Formalization and Preliminary Evaluation of a Pipeline for Text Extraction From Infographics /slideshow/formalization-and-preliminary-evaluation-of-a-pipeline-for-text-extraction-from-infographics/53980556 2015-kdmlv3-151015152532-lva1-app6892
We propose a pipeline for text extraction from infographics that makes use of a novel combination of data mining and computer vision techniques. The pipeline defines a sequence of steps to identify characters, cluster them into text lines, determine their rotation angle, and apply state-of-the-art OCR to recognize the text. In this paper, we formally define the pipeline and present its current implementation. In addition, we have conducted preliminary evaluations over a data corpus of 121 manually annotated infographics from a broad range of illustration types such as bar charts, pie charts, and line charts, maps, and others. We assess the results of our text extraction pipeline by comparing it with two baselines. Finally, we sketch an outline for future work and possibilities for improving the pipeline. - http://ceur-ws.org/Vol-1458/]]>
We propose a pipeline for text extraction from infographics that makes use of a novel combination of data mining and computer vision techniques. The pipeline defines a sequence of steps to identify characters, cluster them into text lines, determine their rotation angle, and apply state-of-the-art OCR to recognize the text. In this paper, we formally define the pipeline and present its current implementation. In addition, we have conducted preliminary evaluations over a data corpus of 121 manually annotated infographics from a broad range of illustration types such as bar charts, pie charts, and line charts, maps, and others. We assess the results of our text extraction pipeline by comparing it with two baselines. Finally, we sketch an outline for future work and possibilities for improving the pipeline. - http://ceur-ws.org/Vol-1458/]]> Thu, 15 Oct 2015 15:25:32 GMT /slideshow/formalization-and-preliminary-evaluation-of-a-pipeline-for-text-extraction-from-infographics/53980556 ascherp@slideshare.net(ascherp) Formalization and Preliminary Evaluation of a Pipeline for Text Extraction From Infographics ascherp We propose a pipeline for text extraction from infographics that makes use of a novel combination of data mining and computer vision techniques. The pipeline defines a sequence of steps to identify characters, cluster them into text lines, determine their rotation angle, and apply state-of-the-art OCR to recognize the text. In this paper, we formally define the pipeline and present its current implementation. In addition, we have conducted preliminary evaluations over a data corpus of 121 manually annotated infographics from a broad range of illustration types such as bar charts, pie charts, and line charts, maps, and others. We assess the results of our text extraction pipeline by comparing it with two baselines. Finally, we sketch an outline for future work and possibilities for improving the pipeline. - http://ceur-ws.org/Vol-1458/ <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/2015-kdmlv3-151015152532-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds" /> We propose a pipeline for text extraction from infographics that makes use of a novel combination of data mining and computer vision techniques. The pipeline defines a sequence of steps to identify characters, cluster them into text lines, determine their rotation angle, and apply state-of-the-art OCR to recognize the text. In this paper, we formally define the pipeline and present its current implementation. In addition, we have conducted preliminary evaluations over a data corpus of 121 manually annotated infographics from a broad range of illustration types such as bar charts, pie charts, and line charts, maps, and others. We assess the results of our text extraction pipeline by comparing it with two baselines. Finally, we sketch an outline for future work and possibilities for improving the pipeline. - http://ceur-ws.org/Vol-1458/

Formalization and Preliminary Evaluation of a Pipeline for Text Extraction From Infographics from Ansgar Scherp

]]> 694 6 https://cdn.slidesharecdn.com/ss_thumbnails/2015-kdmlv3-151015152532-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 A Framework for Iterative Signing of Graph Data on the Web /slideshow/a-framework-for-iterative-signing-of-graph-data-on-the-web/35291362 slides-signinggraphs-final-140530012249-phpapp02
Existing algorithms for signing graph data typically do not cover the whole signing process. In addition, they lack distinctive features such as signing graph data at different levels of granularity, iterative signing of graph data, and signing multiple graphs. In this paper, we introduce a novel framework for signing arbitrary graph data provided, e g., as RDF(S), Named Graphs, or OWL. We conduct an extensive theoretical and empirical analysis of the runtime and space complexity of different framework configurations. The experiments are performed on synthetic and real-world graph data of different size and different number of blank nodes. We investigate security issues, present a trust model, and discuss practical considerations for using our signing framework. We released a Java-based open source implementation of our software framework for iterative signing of arbitrary graph data provided, e. g., as RDF(S), Named Graphs, or OWL. The software framework is based on a formalization of different graph signing functions and supports different configurations. It is available in source code as well as pre-compiled as .jar-file. The graph signing framework exhibits the following unique features: - Signing graphs on different levels of granularity - Signing multiple graphs at once - Iterative signing of graph data for provenance tracking - Independence of the used language for encoding the graph (i. e., the signature does not break when changing the graph representation) The documentation of the software framework and its source code is available from: http://icp.it-risk.iwvi.uni-koblenz.de/wiki/Software_Framework_for_Signing_Graph_Data]]>
Existing algorithms for signing graph data typically do not cover the whole signing process. In addition, they lack distinctive features such as signing graph data at different levels of granularity, iterative signing of graph data, and signing multiple graphs. In this paper, we introduce a novel framework for signing arbitrary graph data provided, e g., as RDF(S), Named Graphs, or OWL. We conduct an extensive theoretical and empirical analysis of the runtime and space complexity of different framework configurations. The experiments are performed on synthetic and real-world graph data of different size and different number of blank nodes. We investigate security issues, present a trust model, and discuss practical considerations for using our signing framework. We released a Java-based open source implementation of our software framework for iterative signing of arbitrary graph data provided, e. g., as RDF(S), Named Graphs, or OWL. The software framework is based on a formalization of different graph signing functions and supports different configurations. It is available in source code as well as pre-compiled as .jar-file. The graph signing framework exhibits the following unique features: - Signing graphs on different levels of granularity - Signing multiple graphs at once - Iterative signing of graph data for provenance tracking - Independence of the used language for encoding the graph (i. e., the signature does not break when changing the graph representation) The documentation of the software framework and its source code is available from: http://icp.it-risk.iwvi.uni-koblenz.de/wiki/Software_Framework_for_Signing_Graph_Data]]> Fri, 30 May 2014 01:22:49 GMT /slideshow/a-framework-for-iterative-signing-of-graph-data-on-the-web/35291362 ascherp@slideshare.net(ascherp) A Framework for Iterative Signing of Graph Data on the Web ascherp Existing algorithms for signing graph data typically do not cover the whole signing process. In addition, they lack distinctive features such as signing graph data at different levels of granularity, iterative signing of graph data, and signing multiple graphs. In this paper, we introduce a novel framework for signing arbitrary graph data provided, e g., as RDF(S), Named Graphs, or OWL. We conduct an extensive theoretical and empirical analysis of the runtime and space complexity of different framework configurations. The experiments are performed on synthetic and real-world graph data of different size and different number of blank nodes. We investigate security issues, present a trust model, and discuss practical considerations for using our signing framework. We released a Java-based open source implementation of our software framework for iterative signing of arbitrary graph data provided, e. g., as RDF(S), Named Graphs, or OWL. The software framework is based on a formalization of different graph signing functions and supports different configurations. It is available in source code as well as pre-compiled as .jar-file. The graph signing framework exhibits the following unique features: - Signing graphs on different levels of granularity - Signing multiple graphs at once - Iterative signing of graph data for provenance tracking - Independence of the used language for encoding the graph (i. e., the signature does not break when changing the graph representation) The documentation of the software framework and its source code is available from: http://icp.it-risk.iwvi.uni-koblenz.de/wiki/Software_Framework_for_Signing_Graph_Data <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/slides-signinggraphs-final-140530012249-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Existing algorithms for signing graph data typically do not cover the whole signing process. In addition, they lack distinctive features such as signing graph data at different levels of granularity, iterative signing of graph data, and signing multiple graphs. In this paper, we introduce a novel framework for signing arbitrary graph data provided, e g., as RDF(S), Named Graphs, or OWL. We conduct an extensive theoretical and empirical analysis of the runtime and space complexity of different framework configurations. The experiments are performed on synthetic and real-world graph data of different size and different number of blank nodes. We investigate security issues, present a trust model, and discuss practical considerations for using our signing framework. We released a Java-based open source implementation of our software framework for iterative signing of arbitrary graph data provided, e. g., as RDF(S), Named Graphs, or OWL. The software framework is based on a formalization of different graph signing functions and supports different configurations. It is available in source code as well as pre-compiled as .jar-file. The graph signing framework exhibits the following unique features: - Signing graphs on different levels of granularity - Signing multiple graphs at once - Iterative signing of graph data for provenance tracking - Independence of the used language for encoding the graph (i. e., the signature does not break when changing the graph representation) The documentation of the software framework and its source code is available from: http://icp.it-risk.iwvi.uni-koblenz.de/wiki/Software_Framework_for_Signing_Graph_Data

A Framework for Iterative Signing of Graph Data on the Web from Ansgar Scherp

]]> 725 2 https://cdn.slidesharecdn.com/ss_thumbnails/slides-signinggraphs-final-140530012249-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Smart photo selection: interpret gaze as personal interest /slideshow/talk-chi2014smartphotoselection/34133440 talk-chi2014-smart-photo-selection-140430110019-phpapp02
Manually selecting subsets of photos from large collections in order to present them to friends or colleagues or to print them as photo books can be a tedious task. Today, fully automatic approaches are at hand for supporting users. They make use of pixel information extracted from the images, analyze contextual information such as capture time and focal aperture, or use both to determine a proper subset of photos. However, these approaches miss the most important factor in the photo selection process: the user. The goal of our approach is to consider individual interests. By recording and analyzing gaze information from the user's viewing photo collections, we obtain information on user's interests and use this information in the creation of personal photo selections. In a controlled experiment with 33 participants, we show that the selections can be significantly improved over a baseline approach by up to 22% when taking individual viewing behavior into account. We also obtained significantly better results for photos taken at an event participants were involved in compared with photos from another event.]]>
Manually selecting subsets of photos from large collections in order to present them to friends or colleagues or to print them as photo books can be a tedious task. Today, fully automatic approaches are at hand for supporting users. They make use of pixel information extracted from the images, analyze contextual information such as capture time and focal aperture, or use both to determine a proper subset of photos. However, these approaches miss the most important factor in the photo selection process: the user. The goal of our approach is to consider individual interests. By recording and analyzing gaze information from the user's viewing photo collections, we obtain information on user's interests and use this information in the creation of personal photo selections. In a controlled experiment with 33 participants, we show that the selections can be significantly improved over a baseline approach by up to 22% when taking individual viewing behavior into account. We also obtained significantly better results for photos taken at an event participants were involved in compared with photos from another event.]]> Wed, 30 Apr 2014 11:00:19 GMT /slideshow/talk-chi2014smartphotoselection/34133440 ascherp@slideshare.net(ascherp) Smart photo selection: interpret gaze as personal interest ascherp Manually selecting subsets of photos from large collections in order to present them to friends or colleagues or to print them as photo books can be a tedious task. Today, fully automatic approaches are at hand for supporting users. They make use of pixel information extracted from the images, analyze contextual information such as capture time and focal aperture, or use both to determine a proper subset of photos. However, these approaches miss the most important factor in the photo selection process: the user. The goal of our approach is to consider individual interests. By recording and analyzing gaze information from the user's viewing photo collections, we obtain information on user's interests and use this information in the creation of personal photo selections. In a controlled experiment with 33 participants, we show that the selections can be significantly improved over a baseline approach by up to 22% when taking individual viewing behavior into account. We also obtained significantly better results for photos taken at an event participants were involved in compared with photos from another event. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/talk-chi2014-smart-photo-selection-140430110019-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Manually selecting subsets of photos from large collections in order to present them to friends or colleagues or to print them as photo books can be a tedious task. Today, fully automatic approaches are at hand for supporting users. They make use of pixel information extracted from the images, analyze contextual information such as capture time and focal aperture, or use both to determine a proper subset of photos. However, these approaches miss the most important factor in the photo selection process: the user. The goal of our approach is to consider individual interests. By recording and analyzing gaze information from the user's viewing photo collections, we obtain information on user's interests and use this information in the creation of personal photo selections. In a controlled experiment with 33 participants, we show that the selections can be significantly improved over a baseline approach by up to 22% when taking individual viewing behavior into account. We also obtained significantly better results for photos taken at an event participants were involved in compared with photos from another event.

Smart photo selection: interpret gaze as personal interest from Ansgar Scherp

]]> 605 2 https://cdn.slidesharecdn.com/ss_thumbnails/talk-chi2014-smart-photo-selection-140430110019-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Events in Multimedia - Theory, Model, Application /slideshow/slideshare-events-in-multimedia-theory-model-application-v07/33651442 slideshare-eventsinmultimedia-theorymodelapplicationv07-140417111140-phpapp02
Talk by Ansgar Scherp. Title: Events in Multimedia - Theory, Model, Application Event: Workshop on Event-based Media Integration and Processing, ACM Multimedia, 2013]]>
Talk by Ansgar Scherp. Title: Events in Multimedia - Theory, Model, Application Event: Workshop on Event-based Media Integration and Processing, ACM Multimedia, 2013]]> Thu, 17 Apr 2014 11:11:40 GMT /slideshow/slideshare-events-in-multimedia-theory-model-application-v07/33651442 ascherp@slideshare.net(ascherp) Events in Multimedia - Theory, Model, Application ascherp Talk by Ansgar Scherp. Title: Events in Multimedia - Theory, Model, Application Event: Workshop on Event-based Media Integration and Processing, ACM Multimedia, 2013 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/slideshare-eventsinmultimedia-theorymodelapplicationv07-140417111140-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Talk by Ansgar Scherp. Title: Events in Multimedia - Theory, Model, Application Event: Workshop on Event-based Media Integration and Processing, ACM Multimedia, 2013

Events in Multimedia - Theory, Model, Application from Ansgar Scherp

]]> 2103 8 https://cdn.slidesharecdn.com/ss_thumbnails/slideshare-eventsinmultimedia-theorymodelapplicationv07-140417111140-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Can you see it? Annotating Image Regions based on Users' Gaze Information /slideshow/can-you-see-it-annotating-image-regions-based-on-users-gaze-information/14801448 eyetracking-basedannotationofimageregions-121019090254-phpapp01
Presentation on eyetracking-based annotation of image regions that I gave at Vienna on Oct 19, 2012. Download original PowerPoint file to enjoy all animations. For the papers, please refer to: http://www.ansgarscherp.net/publications]]>
Presentation on eyetracking-based annotation of image regions that I gave at Vienna on Oct 19, 2012. Download original PowerPoint file to enjoy all animations. For the papers, please refer to: http://www.ansgarscherp.net/publications]]> Fri, 19 Oct 2012 09:02:52 GMT /slideshow/can-you-see-it-annotating-image-regions-based-on-users-gaze-information/14801448 ascherp@slideshare.net(ascherp) Can you see it? Annotating Image Regions based on Users' Gaze Information ascherp Presentation on eyetracking-based annotation of image regions that I gave at Vienna on Oct 19, 2012. Download original PowerPoint file to enjoy all animations. For the papers, please refer to: http://www.ansgarscherp.net/publications <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/eyetracking-basedannotationofimageregions-121019090254-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds" /> Presentation on eyetracking-based annotation of image regions that I gave at Vienna on Oct 19, 2012. Download original PowerPoint file to enjoy all animations. For the papers, please refer to: http://www.ansgarscherp.net/publications

Can you see it? Annotating Image Regions based on Users' Gaze Information from Ansgar Scherp

]]> 1347 4 https://cdn.slidesharecdn.com/ss_thumbnails/eyetracking-basedannotationofimageregions-121019090254-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Linked open data - how to juggle with more than a billion triples /slideshow/linked-open-data-how-to-juggle-with-more-than-a-billion-triples/14763272 linkedopendata-howtojugglewithmorethanabilliontriples-121017041912-phpapp02
�ݺ�ߣs of my inauguration talk at the University of Mannheim in Germany in October 2012. Download this slide set to enjoy all animations. ]]>
�ݺ�ߣs of my inauguration talk at the University of Mannheim in Germany in October 2012. Download this slide set to enjoy all animations. ]]> Wed, 17 Oct 2012 04:19:10 GMT /slideshow/linked-open-data-how-to-juggle-with-more-than-a-billion-triples/14763272 ascherp@slideshare.net(ascherp) Linked open data - how to juggle with more than a billion triples ascherp �ݺ�ߣs of my inauguration talk at the University of Mannheim in Germany in October 2012. Download this slide set to enjoy all animations. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/linkedopendata-howtojugglewithmorethanabilliontriples-121017041912-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs of my inauguration talk at the University of Mannheim in Germany in October 2012. Download this slide set to enjoy all animations.

Linked open data - how to juggle with more than a billion triples from Ansgar Scherp

]]> 1395 2 https://cdn.slidesharecdn.com/ss_thumbnails/linkedopendata-howtojugglewithmorethanabilliontriples-121017041912-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SchemEX -- Building an Index for Linked Open Data /slideshow/lod-introschem-exoslov3-13951803/13951803 lod-introschemex-oslo-v3-120812164815-phpapp01
General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations.]]>
General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations.]]> Sun, 12 Aug 2012 16:48:14 GMT /slideshow/lod-introschem-exoslov3-13951803/13951803 ascherp@slideshare.net(ascherp) SchemEX -- Building an Index for Linked Open Data ascherp General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/lod-introschemex-oslo-v3-120812164815-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds" /> General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations.

SchemEX -- Building an Index for Linked Open Data from Ansgar Scherp

]]> 2157 3 https://cdn.slidesharecdn.com/ss_thumbnails/lod-introschemex-oslo-v3-120812164815-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SchemEX -- Building an Index for Linked Open Data /slideshow/lod-introschem-exoslov3/13949812 lod-introschemex-oslo-v3-120812090304-phpapp01
General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations.]]>
General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations.]]> Sun, 12 Aug 2012 09:03:02 GMT /slideshow/lod-introschem-exoslov3/13949812 ascherp@slideshare.net(ascherp) SchemEX -- Building an Index for Linked Open Data ascherp General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/lod-introschemex-oslo-v3-120812090304-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds" /> General introduction to Linked Open Data and schema extraction using SchemEX. Download full slide set to enjoy all animations.

SchemEX -- Building an Index for Linked Open Data from Ansgar Scherp

]]> 551 3 https://cdn.slidesharecdn.com/ss_thumbnails/lod-introschemex-oslo-v3-120812090304-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 A Model of Events for Integrating Event-based Information in Complex Socio-technical Information Spaces /slideshow/a-model-of-events/11737586 scherpetal-amodelofevents-120224124927-phpapp01
�ݺ�ߣs about our core ontology Event-Model-F. Download slides to enjoy all animations.]]>
�ݺ�ߣs about our core ontology Event-Model-F. Download slides to enjoy all animations.]]> Fri, 24 Feb 2012 12:49:24 GMT /slideshow/a-model-of-events/11737586 ascherp@slideshare.net(ascherp) A Model of Events for Integrating Event-based Information in Complex Socio-technical Information Spaces ascherp �ݺ�ߣs about our core ontology Event-Model-F. Download slides to enjoy all animations. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/scherpetal-amodelofevents-120224124927-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs about our core ontology Event-Model-F. Download slides to enjoy all animations.

A Model of Events for Integrating Event-based Information in Complex Socio-technical Information Spaces from Ansgar Scherp

]]> 994 3 https://cdn.slidesharecdn.com/ss_thumbnails/scherpetal-amodelofevents-120224124927-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SchemEX - Creating the Yellow Pages for the Linked Open Data Cloud /slideshow/schemex-creating-the-yellow-pages-for-the-linked-open-data-cloud/11737502 konrathgottronscherp-schemex-120224124256-phpapp01
�ݺ�ߣs of the billion triple challenge 2011 on SchemEX. Please download original file to enjoy all animations.]]>
�ݺ�ߣs of the billion triple challenge 2011 on SchemEX. Please download original file to enjoy all animations.]]> Fri, 24 Feb 2012 12:42:53 GMT /slideshow/schemex-creating-the-yellow-pages-for-the-linked-open-data-cloud/11737502 ascherp@slideshare.net(ascherp) SchemEX - Creating the Yellow Pages for the Linked Open Data Cloud ascherp �ݺ�ߣs of the billion triple challenge 2011 on SchemEX. Please download original file to enjoy all animations. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/konrathgottronscherp-schemex-120224124256-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs of the billion triple challenge 2011 on SchemEX. Please download original file to enjoy all animations.

SchemEX - Creating the Yellow Pages for the Linked Open Data Cloud from Ansgar Scherp

]]> 1101 2 https://cdn.slidesharecdn.com/ss_thumbnails/konrathgottronscherp-schemex-120224124256-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 strukt - A Pattern System for Integrating Individual and Organizational Knowledge Work /slideshow/scherp-struktpresentationiswc/11368183 scherp-strukt-presentation-iswc-120201082339-phpapp02
Please download these slides to enjoy all animations,]]>
Please download these slides to enjoy all animations,]]> Wed, 01 Feb 2012 08:23:36 GMT /slideshow/scherp-struktpresentationiswc/11368183 ascherp@slideshare.net(ascherp) strukt - A Pattern System for Integrating Individual and Organizational Knowledge Work ascherp Please download these slides to enjoy all animations, <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/scherp-strukt-presentation-iswc-120201082339-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Please download these slides to enjoy all animations,

strukt - A Pattern System for Integrating Individual and Organizational Knowledge Work from Ansgar Scherp

]]> 509 3 https://cdn.slidesharecdn.com/ss_thumbnails/scherp-strukt-presentation-iswc-120201082339-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Identifying Objects in Images from Analyzing the User‘s Gaze Movements for Provided Tags /slideshow/identifying-objects-in-images-from-analyzing-the-users-gaze-movements-for-provided-tags/10820002 mmm2012-eyetracking-meets-labelme-120105093319-phpapp02
�ݺ�ߣs of our MMM 2012 paper. Download slides to enjoy all animations.]]>
�ݺ�ߣs of our MMM 2012 paper. Download slides to enjoy all animations.]]> Thu, 05 Jan 2012 09:33:16 GMT /slideshow/identifying-objects-in-images-from-analyzing-the-users-gaze-movements-for-provided-tags/10820002 ascherp@slideshare.net(ascherp) Identifying Objects in Images from Analyzing the User‘s Gaze Movements for Provided Tags ascherp �ݺ�ߣs of our MMM 2012 paper. Download slides to enjoy all animations. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/mmm2012-eyetracking-meets-labelme-120105093319-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs of our MMM 2012 paper. Download slides to enjoy all animations.

Identifying Objects in Images from Analyzing the User‘s Gaze Movements for Provided Tags from Ansgar Scherp

]]> 1318 2 https://cdn.slidesharecdn.com/ss_thumbnails/mmm2012-eyetracking-meets-labelme-120105093319-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 https://cdn.slidesharecdn.com/profile-photo-ascherp-48x48.jpg?cb=1731505990 https://cdn.slidesharecdn.com/ss_thumbnails/graphsumpresentation02final-220124212001-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/analysis-of-graphsums-attention-weights-to-improve-the-explainability-of-multidocument-summarization/251047624 Analysis of GraphSum's... https://cdn.slidesharecdn.com/ss_thumbnails/stereoiiwas-presentation-220124211535-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/stereo-a-pipeline-for-extracting-experiment-statistics-conditions-and-topics-from-scientific-papers/251047618 STEREO: A Pipeline for... https://cdn.slidesharecdn.com/ss_thumbnails/jessenboeschenscherp-nntextextraction-190926130309-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/text-localization-in-scientific-figures-using-fully-convolutional-neural-networks-on-limited-training-data/176331413 Text Localization in S...