Kluegl, P.; Toepfer, M.; Lemmerich, F.; Hotho, A. & Puppe, F.: Exploiting Structural Consistencies with Stacked Conditional Random Fields. In: Mathematical Methodologies in Pattern Recognition and Machine Learning Springer Proceedings in Mathematics & Statistics 30 (2013), S. 111-125
Conditional Random Fields (CRF) are popular methods for labeling unstructured or textual data. Like many machine learning approaches, these undirected graphical models assume the instances to be independently distributed. However, in real-world applications data is grouped in a natural way, e.g., by its creation context. The instances in each group often share additional structural consistencies. This paper proposes a domain-independent method for exploiting these consistencies by combining two CRFs in a stacked learning framework. We apply rule learning collectively on the predictions of an initial CRF for one context to acquire descriptions of its specific properties. Then, we utilize these descriptions as dynamic and high quality features in an additional (stacked) CRF. The presented approach is evaluated with a real-world dataset for the segmentation of references and achieves a significant reduction of the labeling error.
@article{kluegl2013exploiting,
author = {Kluegl, Peter and Toepfer, Martin and Lemmerich, Florian and Hotho, Andreas and Puppe, Frank},
title = {Exploiting Structural Consistencies with Stacked Conditional Random Fields},
journal = {Mathematical Methodologies in Pattern Recognition and Machine Learning Springer Proceedings in Mathematics & Statistics},
year = {2013},
volume = {30},
pages = {111-125},
keywords = {ie, learning, 2013, myown, references},
abstract = {Conditional Random Fields (CRF) are popular methods for labeling unstructured or textual data. Like many machine learning approaches, these undirected graphical models assume the instances to be independently distributed. However, in real-world applications data is grouped in a natural way, e.g., by its creation context. The instances in each group often share additional structural consistencies. This paper proposes a domain-independent method for exploiting these consistencies by combining two CRFs in a stacked learning framework. We apply rule learning collectively on the predictions of an initial CRF for one context to acquire descriptions of its specific properties. Then, we utilize these descriptions as dynamic and high quality features in an additional (stacked) CRF. The presented approach is evaluated with a real-world dataset for the segmentation of references and achieves a significant reduction of the labeling error.}
}
Atzmueller, M.; Beer, S. & Puppe, F.: Data Mining, Validation and Collaborative Knowledge Capture. In: Brüggemann, S. & d’Amato, C. (Hrsg.): Collaboration and the Semantic Web: Social Networks, Knowledge Networks, and Knowledge Resources. IGI Global, 2012, S. 149-167
@incollection{ABP:11,
author = {Atzmueller, Martin and Beer, Stephanie and Puppe, Frank},
title = {Data Mining, Validation and Collaborative Knowledge Capture},
editor = {Brüggemann, Stefan and d’Amato, Claudia},
booktitle = {Collaboration and the Semantic Web: Social Networks, Knowledge Networks, and Knowledge Resources},
publisher = {IGI Global},
year = {2012},
pages = {149-167},
keywords = {textmarker, nlp, 2012, data, mining}
}
Klügl, P.; Toepfer, M.; Lemmerich, F.; Hotho, A. & Puppe, F.: Collective Information Extraction with Context-Specific Consistencies.. In: Flach, P. A.; Bie, T. D. & Cristianini, N. (Hrsg.): ECML/PKDD (1). Springer, 2012 (Lecture Notes in Computer Science 7523), S. 728-743
[Volltext]
@inproceedings{conf/pkdd/KluglTLHP12,
author = {Klügl, Peter and Toepfer, Martin and Lemmerich, Florian and Hotho, Andreas and Puppe, Frank},
title = {Collective Information Extraction with Context-Specific Consistencies.},
editor = {Flach, Peter A. and Bie, Tijl De and Cristianini, Nello},
booktitle = {ECML/PKDD (1)},
series = {Lecture Notes in Computer Science},
publisher = {Springer},
year = {2012},
volume = {7523},
pages = {728-743},
url = {http://dblp.uni-trier.de/db/conf/pkdd/pkdd2012-1.html#KluglTLHP12},
isbn = {978-3-642-33459-7},
keywords = {information, ie, myown, 2012, extraction, context}
}
Klügl, P.; Toepfer, M.; Lemmerich, F.; Hotho, A. & Puppe, F.: Stacked Conditional Random Fields Exploiting Structural Consistencies. In: Carmona, P. L.; Sánchez, J. S. & Fred, A. (Hrsg.): Proceedings of 1st International Conference on Pattern Recognition Applications and Methods ICPRAM. Vilamoura, Algarve, Portugal: SciTePress, 2012, S. 240-248
[Volltext]
Conditional Random Fields CRF are popular methods for labeling unstructured or textual data. Like many machine learning approaches these undirected graphical models assume the instances to be independently distributed. However, in real world applications data is grouped in a natural way, e.g., by its creation context. The instances in each group often share additional structural consistencies. This paper proposes a domain-independent method for exploiting these consistencies by combining two CRFs in a stacked learning framework. The approach incorporates three successive steps of inference: First, an initial CRF processes single instances as usual. Next, we apply rule learning collectively on all labeled outputs of one context to acquire descriptions of its specific properties. Finally, we utilize these descriptions as dynamic and high quality features in an additional stacked CRF. The presented approach is evaluated with a real-world dataset for the segmentation of references and achieves a significant reduction of the labeling error.
@inproceedings{kluegl2012stacked,
author = {Klügl, Peter and Toepfer, Martin and Lemmerich, Florian and Hotho, Andreas and Puppe, Frank},
title = {Stacked Conditional Random Fields Exploiting Structural Consistencies},
editor = {Carmona, Pedro Latorre and Sánchez, J. Salvador and Fred, Ana},
booktitle = {Proceedings of 1st International Conference on Pattern Recognition Applications and Methods ICPRAM},
publisher = {SciTePress},
address = {Vilamoura, Algarve, Portugal},
year = {2012},
pages = {240-248},
url = {http://ki.informatik.uni-wuerzburg.de/papers/pkluegl/2012-ICPRAM-StackedCRF.pdf},
keywords = {crf, stacked, fields, conditional, myown, 2012, random},
abstract = {Conditional Random Fields CRF are popular methods for labeling unstructured or textual data. Like many machine learning approaches these undirected graphical models assume the instances to be independently distributed. However, in real world applications data is grouped in a natural way, e.g., by its creation context. The instances in each group often share additional structural consistencies. This paper proposes a domain-independent method for exploiting these consistencies by combining two CRFs in a stacked learning framework. The approach incorporates three successive steps of inference: First, an initial CRF processes single instances as usual. Next, we apply rule learning collectively on all labeled outputs of one context to acquire descriptions of its specific properties. Finally, we utilize these descriptions as dynamic and high quality features in an additional stacked CRF. The presented approach is evaluated with a real-world dataset for the segmentation of references and achieves a significant reduction of the labeling error.}
}
Atzmueller, M.; Beer, S. & Puppe, F.: Data Mining, Validation and Collaborative Knowledge Capture. In: Brüggemann, S. & d’Amato, C. (Hrsg.): Collaboration and the Semantic Web: Social Networks, Knowledge Networks and Knowledge Resources.. IGI Global, 2011
@incollection{ABP:11,
author = {Atzmueller, Martin and Beer, Stephanie and Puppe, Frank},
title = {Data Mining, Validation and Collaborative Knowledge Capture},
editor = {Brüggemann, Stefan and d’Amato, Claudia},
booktitle = {Collaboration and the Semantic Web: Social Networks, Knowledge Networks and Knowledge Resources.},
publisher = {IGI Global},
year = {2011},
keywords = {textmarker, itegpub, nlp, 2011, myown, data, mining}
}
Toepfer, M.; Kluegl, P.; Hotho, A. & Puppe, F.: Segmentation of References with Skip-Chain Conditional Random Fields for Consistent Label Transitions. Workshop Notes of the LWA 2011 - Learning, Knowledge, Adaptation. 2011
[Volltext]
@inproceedings{toepfer2011segmentation,
author = {Toepfer, Martin and Kluegl, Peter and Hotho, Andreas and Puppe, Frank},
title = {Segmentation of References with Skip-Chain Conditional Random Fields for Consistent Label Transitions},
booktitle = {Workshop Notes of the LWA 2011 - Learning, Knowledge, Adaptation},
year = {2011},
url = {http://ki.informatik.uni-wuerzburg.de/papers/pkluegl/2011-LWA-SkYp.pdf},
keywords = {chain, 2011, conditional, myown, segmentation, references}
}
Kluegl, P.; Hotho, A. & Puppe, F.: Local Adaptive Extraction of References. In: Dillmann, R.; Beyerer, J.; Hanebeck, U. D. & Schultz, T. (Hrsg.): KI 2010: Advances in Artificial Intelligence, 33rd Annual German Conference on AI. Springer, 2010 LNAI 6359 , S. 40-47
[Volltext]
The accurate extraction of scholarly reference information from scientific publications is essential for many useful applications like BibTeX management systems or citation analysis. Automatic extraction methods suffer from the heterogeneity of reference notation, no matter wether the extraction model was handcrafted or learnt from labeled data. However, references of the same paper or journal are usually homogeneous. We exploit this local consistency with a novel approach. Given some initial information from such a reference section, we try to derived generalized patterns. These patterns are used to create a local model of the current document. The local model helps to identify errors and to improve the extracted information incrementally during the extraction process. Our approach is implemented with handcrafted transformation rules working on a meta-level being able to correct the information independent of the applied layout style. The experimental results compete very well with the state of the art methods and show an extremely high performance on consistent reference sections.
@inproceedings{2010-KI-KHP,
author = {Kluegl, Peter and Hotho, Andreas and Puppe, Frank},
title = {Local Adaptive Extraction of References},
editor = {Dillmann, Rüdiger and Beyerer, Jürgen and Hanebeck, Uwe D. and Schultz, Tanja},
booktitle = {KI 2010: Advances in Artificial Intelligence, 33rd Annual German Conference on AI},
series = { LNAI 6359},
publisher = {Springer},
year = {2010},
pages = {40-47},
url = {http://ki.informatik.uni-wuerzburg.de/papers/pkluegl/2010-KI-LAER.pdf},
isbn = {978-3-642-16110-0},
keywords = {ie, 2010, myown, scholary, references},
abstract = {The accurate extraction of scholarly reference information from scientific publications is essential for many useful applications like BibTeX management systems or citation analysis. Automatic extraction methods suffer from the heterogeneity of reference notation, no matter wether the extraction model was handcrafted or learnt from labeled data. However, references of the same paper or journal are usually homogeneous. We exploit this local consistency with a novel approach. Given some initial information from such a reference section, we try to derived generalized patterns. These patterns are used to create a local model of the current document. The local model helps to identify errors and to improve the extracted information incrementally during the extraction process. Our approach is implemented with handcrafted transformation rules working on a meta-level being able to correct the information independent of the applied layout style. The experimental results compete very well with the state of the art methods and show an extremely high performance on consistent reference sections. }
}
Toepfer, M.; Kluegl, P.; Hotho, A. & Puppe., F.: Conditional Random Fields For Local Adaptive Reference Extraction. In: Atzmüller, M.; Benz, D.; Hotho, A. & Stumme, G. (Hrsg.): Proceedings of LWA2010 - Workshop-Woche: Lernen, Wissen & Adaptivitaet. Kassel, Germany: 2010
[Volltext]
The accurate extraction of bibliographic information from scientific publications is an active field of research. Machine learning and sequence labeling approaches like Conditional Random Fields (CRF) are often applied for this reference extraction task, but still suffer from the ambiguity of reference notation. Reference sections apply a predefined style guide and contain only homogeneous references. Therefore, other references of the same paper or journal often provide evidence how the fields of a reference are correctly labeled. We propose a novel approach that exploits the similarities within a document. Our process model uses information of unlabeled documents directly during the extraction task in order to automatically adapt to the perceived style guide. This is implemented by changing the manifestation of the features for the applied CRF. The experimental results show considerable improvements compared to the common approach. We achieve an average F1 score of 96.7% and an instance accuracy of 85.4% on the test data set.
@inproceedings{kdml21,
author = {Toepfer, Martin and Kluegl, Peter and Hotho, Andreas and Puppe., Frank},
title = {Conditional Random Fields For Local Adaptive Reference Extraction},
editor = {Atzmüller, Martin and Benz, Dominik and Hotho, Andreas and Stumme, Gerd},
booktitle = {Proceedings of LWA2010 - Workshop-Woche: Lernen, Wissen & Adaptivitaet},
address = {Kassel, Germany},
year = {2010},
url = {http://www.kde.cs.uni-kassel.de/conf/lwa10/papers/kdml21.pdf},
keywords = {information, crf, 2010, myown, extraction},
abstract = {The accurate extraction of bibliographic information from scientific publications is an active field of research. Machine learning and sequence labeling approaches like Conditional Random Fields (CRF) are often applied for this reference extraction task, but still suffer from the ambiguity of reference notation. Reference sections apply a predefined style guide and contain only homogeneous references. Therefore, other references of the same paper or journal often provide evidence how the fields of a reference are correctly labeled. We propose a novel approach that exploits the similarities within a document. Our process model uses information of unlabeled documents directly during the extraction task in order to automatically adapt to the perceived style guide. This is implemented by changing the manifestation of the features for the applied CRF. The experimental results show considerable improvements compared to the common approach. We achieve an average F1 score of 96.7% and an instance accuracy of 85.4% on the test data set.}
}
Atzmueller, M.; Beer, S. & Puppe, F.: A Data Warehouse-Based Approach for Quality Management, Evaluation and Analysis of Intelligent Systems using Subgroup Mining. Proc. 22nd International Florida Artificial Intelligence Research Society Conference (FLAIRS), accepted. AAAI Press, 2009, S. 372-377
@inproceedings{ABP:09,
author = {Atzmueller, Martin and Beer, Stephanie and Puppe, Frank},
title = {A Data Warehouse-Based Approach for Quality Management, Evaluation and Analysis of Intelligent Systems using Subgroup Mining},
booktitle = {Proc. 22nd International Florida Artificial Intelligence Research Society Conference (FLAIRS), accepted},
publisher = {AAAI Press},
year = {2009},
pages = {372-377},
keywords = {discovery, experience, semantic, myown, analytics, data, visual, mining, management, knowledge}
}
Atzmueller, M.; Puppe, F. & Buscher, H.-P.: A Semi-Automatic Approach for Confounding-Aware Subgroup Discovery. In: International Journal on Artificial Intelligence Tools (IJAIT) 18 (2009), Nr. 1, S. 1 - 18
@article{APB:09,
author = {Atzmueller, Martin and Puppe, Frank and Buscher, Hans-Peter},
title = {A Semi-Automatic Approach for Confounding-Aware Subgroup Discovery},
journal = {International Journal on Artificial Intelligence Tools (IJAIT)},
year = {2009},
volume = {18},
number = {1},
pages = {1 -- 18},
keywords = {discovery, experience, semantic, myown, analytics, data, visual, mining, management, knowledge}
}
Atzmueller, M.; Beer, S.; Hörnlein, A.; Melcher, R.; Lührs, H. & Puppe, F.: Design and Implementation of a Data Warehouse for Quality Management, System Evaluation and Knowledge Discovery in the Medical Domain. Proc. 1st European Workshop on Design, Evaluation and Refinement of Intelligent Systems. Erfurt: 2008
@inproceedings{ABHMLP:08,
author = {Atzmueller, Martin and Beer, Stephanie and Hörnlein, Alexander and Melcher, Ralf and Lührs, Hardi and Puppe, Frank},
title = {Design and Implementation of a Data Warehouse for Quality Management, System Evaluation and Knowledge Discovery in the Medical Domain},
booktitle = {Proc. 1st European Workshop on Design, Evaluation and Refinement of Intelligent Systems},
address = {Erfurt},
year = {2008},
keywords = {discovery, experience, semantic, myown, analytics, data, visual, mining, management, warehouse, knowledge}
}
Puppe, F.; Atzmueller, M.; Buscher, G.; Huettig, M.; Lührs, H. & Buscher, H.-P.: Application and Evaluation of a Medical Knowledge-System in Sonography (SonoConsult). Proc. 18th European Conference on Artificial Intelligence (ECAI 20008), accepted. 2008
@inproceedings{PABHLB:08,
author = {Puppe, Frank and Atzmueller, Martin and Buscher, Georg and Huettig, Matthias and Lührs, Hardi and Buscher, Hans-Peter},
title = {Application and Evaluation of a Medical Knowledge-System in Sonography (SonoConsult)},
booktitle = {Proc. 18th European Conference on Artificial Intelligence (ECAI 20008), accepted},
year = {2008},
keywords = {refinement, diagnosis, system, myown, data, mining, intelligent}
}
Atzmueller, M. & Puppe, F.: Causal Subgroup Analysis for Detecting Confounding. Proc. 18th International Conference on Applications of Declarative Programming and Knowledge Management (INAP 2007). Wuerzburg, Germany: 2007
@inproceedings{AP:07a,
author = {Atzmueller, Martin and Puppe, Frank},
title = {Causal Subgroup Analysis for Detecting Confounding},
booktitle = {Proc. 18th International Conference on Applications of Declarative Programming and Knowledge Management (INAP 2007)},
address = {Wuerzburg, Germany},
year = {2007},
keywords = {subgroup-discovery, myown, data, mining}
}
Atzmueller, M.; Baumeister, J.; Klügl, P. & Puppe, F.: Rapid Knowledge Capture Using Subgroup Discovery with Incremental Refinement. Proc. 4th International Conference on Knowledge Capture (K-CAP 2007). ACM Press, 2007, S. 31-38
@inproceedings{ABKP:07,
author = {Atzmueller, Martin and Baumeister, Joachim and Klügl, Peter and Puppe, Frank},
title = {Rapid Knowledge Capture Using Subgroup Discovery with Incremental Refinement},
booktitle = {Proc. 4th International Conference on Knowledge Capture (K-CAP 2007)},
publisher = {ACM Press},
year = {2007},
pages = {31--38},
keywords = {capture, subgroup-discovery, myown, knowledge}
}
Atzmueller, M. & Puppe, F.: Case-Based Characterization and Analysis of Subgroup Patterns. Proc. LWA 2006 (KDML Special Track), Hildesheimer Informatik Berichte. University of Hildesheim, 2006
@inproceedings{AP:06a,
author = {Atzmueller, Martin and Puppe, Frank},
title = {Case-Based Characterization and Analysis of Subgroup Patterns},
booktitle = {Proc. LWA 2006 (KDML Special Track), Hildesheimer Informatik Berichte},
publisher = {University of Hildesheim},
year = {2006},
keywords = {subgroup-discovery, myown, data, mining}
}
Atzmueller, M. & Puppe, F.: SD-Map - A Fast Algorithm for Exhaustive Subgroup Discovery. Proc. 10th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD 2006). 2006LNAI , S. 6-17
@inproceedings{AP:06a,
author = {Atzmueller, Martin and Puppe, Frank},
title = {SD-Map -- A Fast Algorithm for Exhaustive Subgroup Discovery},
booktitle = {Proc. 10th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD 2006)},
series = {LNAI},
year = {2006},
number = {4213},
pages = {6-17},
keywords = {imported, myown}
}
Baumeister, J.; Atzmueller, M.; Kluegl, P. & Puppe, F.: Conservative and Creative Strategies for the Refinement of Scoring Rules. In: Sutcliffe, G. & Goebel, R. (Hrsg.): Proc. 19th Intl. Florida Artificial Intelligence Research Society Conference 2006 (FLAIRS-2006). AAAI Press, 2006, S. 408-413
@inproceedings{BAKP:06,
author = {Baumeister, Joachim and Atzmueller, Martin and Kluegl, Peter and Puppe, Frank},
title = {Conservative and Creative Strategies for the Refinement of Scoring Rules},
editor = {Sutcliffe, Geoff and Goebel, Randy},
booktitle = {Proc. 19th Intl. Florida Artificial Intelligence Research Society Conference 2006 (FLAIRS-2006)},
publisher = {AAAI Press},
year = {2006},
pages = {408--413},
keywords = {refinement, diagnosis, subgroup-discovery, myown}
}
Atzmueller, M.; Baumeister, J. & Puppe, F.: Exemplifying Subgroup Mining Results for Interactive Knowledge Refinement. Proc. 13th Leipziger Informatik-Tage 2005 (LIT 2005). 2005LNI , S. 101-106
@inproceedings{ABP:05,
author = {Atzmueller, Martin and Baumeister, Joachim and Puppe, Frank},
title = {Exemplifying Subgroup Mining Results for Interactive Knowledge Refinement},
booktitle = {Proc. 13th Leipziger Informatik-Tage 2005 (LIT 2005)},
series = {LNI},
year = {2005},
pages = {101-106},
keywords = {discovery, experience, semantic, vikamine, myown, analytics, data, visual, mining, management, knowledge}
}
Atzmueller, M.; Baumeister, J.; Hemsing, A.; Richter, E.-J. & Puppe, F.: Subgroup Mining for Interactive Knowledge Refinement. Proc. 10th Conference on Artificial Intelligence in Medicine (AIME 05). 2005LNAI 3581 , S. 453-462
@inproceedings{ABHRP:05,
author = {Atzmueller, Martin and Baumeister, Joachim and Hemsing, Achim and Richter, Ernst-Jürgen and Puppe, Frank},
title = {Subgroup Mining for Interactive Knowledge Refinement},
booktitle = {Proc. 10th Conference on Artificial Intelligence in Medicine (AIME 05)},
series = {LNAI 3581},
year = {2005},
pages = {453--462},
keywords = {refinement, discovery, experience, semantic, vikamine, myown, analytics, data, visual, mining, management, knowledge}
}
Baumeister, J.; Atzmueller, M. & Puppe, F.: Inductive Learning for Case-Based Diagnosis with Multiple Faults. Advances in Case-Based Reasoning. 2002 (LNAI 2416), S. 28-42
@inproceedings{BAP:02,
author = {Baumeister, Joachim and Atzmueller, Martin and Puppe, Frank},
title = {Inductive Learning for Case-Based Diagnosis with Multiple Faults},
booktitle = {Advances in Case-Based Reasoning},
series = {LNAI},
year = {2002},
volume = {2416},
pages = {28-42},
note = {Proc. 6th European Conference on Case-Based Reasoning (ECCBR 2002)},
keywords = {refinement, diagnosis, discovery, experience, semantic, vikamine, myown, analytics, visual, data, mining, management, knowledge}
}