Davide Magatti, Ph.D.
Since June, 2011 I am a Postdoctoral Reseracher at Politecnico di Milano - DOOR (Data Mining Optimization, Operations Research) Group.
My new home page is Here
Dipartimento di Informatica, Sistemistica e Comunicazione
Università degli Studi di Milano-Bicocca
U14, Viale Sarca 336
20126 Milano, ITALY
magatti@disco.unimib.it
Skype ID: davmago
Twitter: davmago
LinkedIn Profile
I received my Ph.D in Computer Science on February, 8th 2011 at the Department of Informatics, System and Communication (DISCo) - Università degli studi di Milano-Bicocca.
My research is mainly involved in document management and information extraction. According to different studies the information growth in the next ten years will generate a massive information overload that only strong computational approach will be able to deal with. I am interested in real world applications of my research projects.
Research
My research activity is focused on models and algorithms for Text Mining. I am particularly interested in the following tasks:
- Document Clustering and Organization
- Information Integration
- Information Extraction
Main models and algorithms include
- Latent Dirichlet Allocation and its extensions
- Hierarchical Topic extraction and Stochastic Processes (Urn models)
- Conditional Random Fields for sequential tagging and information extraction
Corpora
Here you can find some references to corpora used in Topic Extraction models and Text Mining tasks.
Ph.D Thesis
Davide Magatti. Graphical Models for Text Mining: Knowledge Extraction and Performance Estimation Ph.D Thesis Full Thesis, Slides
Publications
D. Magatti, F. Stella and M. Faini. A Software System for Topic Extraction and Documemnt Classification in Proceeding of 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology PDF
D. Magatti, S.Calegari, D. Ciucci, F. Stella. Automatic labeling of Topics in Proceedings of 2009 IEEE International Conferences on Intelligent Systems Design and Applications (ISDA) PDF
D. Magatti, F. Steinke, M. Bundschus, V. Tresp. Combined Structured and Keyword-Based Search in Textually Enriched Entity-Relationship Graphs Accepted paper at AKBC2010 - First Workshop on Automated Knowledge Base Construction PDF SLIDES VideoLectures.net
Ramirez H. E., Brena R., Magatti D., Stella F. Probabilistic Metrics for Soft-Clustering and Topic Model Validation in Proceeedings of 2010 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology PDF
D. Magatti, F. Stella. Probabilistic topic discovery and Automatic Topic Tagging invited chapter in Quantitative Semantics and Soft Computing Methods for the Web. IGI Global 2011. Corpus and Results
Academic Career
- 2010 (Jan - May): Intern at Siemens AG - Corporate Research München
- 2007 - current: Ph.D. School in Informatics at Disco, Università degli Studi di Milano-Bicocca
- 2005 - 2007: Computer Science Master Degree, Università degli Studi di Milano-Bicocca
- 2001 - 2005: Computer Science Degree, Università degli Studi di Milano-Bicocca
Awards
- Winner of a grant by Telecom Italia in Working Capital Tour for the project: Side Informer - Personalized Information Manager SLIDE, PRESS VideoPresentation (in Italian!), Interview (in Italian)