Dissertation Defense

TABEL — A Domain Independent and Extensible
Framework for Inferring the Semantics of Tables

Varish Vyankatesh Mulwad

8:00am Thursday, 8 January 2015, ITE325b

Tables are an integral part of documents, reports and Web pages in many scientific and technical domains, compactly encoding important information that can be difficult to express in text. Table-like structures outside documents, such as spreadsheets, CSV files, log files and databases, are widely used to represent and share information. However, tables remain beyond the scope of regular text processing systems which often treat them like free text.

This dissertation presents TABEL — a domain independent and extensible framework to infer the semantics of tables and represent them as RDF Linked Data. TABEL captures the intended meaning of a table by mapping header cells to classes, data cell values to existing entities and pair of columns to relations from an given ontology and knowledge base. The core of the framework consists of a module that represents a table as a graphical model to jointly infer the semantics of headers, data cells and relation between headers. We also introduce a novel Semantic Message Passing scheme, which incorporates semantics into message passing, to perform joint inference over the probabilistic graphical model. We also develop and explore a “human-in-the-loop” paradigm, presenting plausible models of user interaction with our framework and its impact on the quality of inferred semantics.

We present techniques that are both extensible and domain agnostic. Our framework supports the addition of preprocessing modules without affecting existing ones, making TABEL extensible. It also allows background knowledge bases to be adapted and changed based on the domains of the tables, thus making it domain independent. We demonstrate the extensibility and domain independence of our techniques by developing an application of TABEL in the healthcare domain. We develop a proof of concept for an application to generate meta-analysis reports automatically, which is built on top of the semantics inferred from tables found in medical literature.

A thorough evaluation with experiments over dataset of tables from the Web and medical research reports presents promising results.

Committee: Drs. Tim Finin (chair), Tim Oates, Anupam Joshi, Yun Peng, Indrajit Bhattacharya (IBM Research) and L. V. Subramaniam (IBM Research)

UMBC GM Niclas “The Dark Knight” Huschenbeth plays GM Nadezhda Kosintseva from UT Dallas in the fifth round of the Pan-Am chess tournament.

Congratulations to the UMBC chess team for placing third in the 2014 Pan American Intercollegiate Team Chess Championship and qualifying to advance to the President’s Cup on April 5-6 at the New York Athletic Club.  This tournament, known as the Final Four of College Chess, will determine the 2015 U.S. college team chess champion.

The UMBC team in this year’s Pan-Am consists of GM Niclas Huschenbeth, IM Levan Bregadze, IM Tanguy Ringoir, GM Akshayraj Kore and WGM Nazi Paikidze (alternate).

The final standings were: Webster University A (5.5 points), the University of Texas at Dallas A (5 points), the University of Maryland, Baltimore County (5 points) and Texas Tech University A (4.5 points). UTD placed ahead of UMBC by virtue of winning their match.

You can view the final results on the tournament website and review selected games on Monroi.

Single Carrot Theatre seeks male actors of color to perform in its upcoming production of Utopia Parkway by Charles Mee and directed by SCT's Artistic Director Genevieve de Mahy. Must be able to carry a tune, and must be willing to work collaboratively. The ideal candidate would be available for most rehearsals and all of tech and performances (pertinent dates listed below). The ability to play one or several musical instruments is a plus.

Dissertation Defense

TABEL — A Domain Independent and Extensible Framework
for Inferring the Semantics of Tables

Varish Vyankatesh Mulwad

8:00am Thursday, 8 January 2015, ITE325b

Tables are an integral part of documents, reports and Web pages in many scientific and technical domains, compactly encoding important information that can be difficult to express in text. Table-like structures outside documents, such as spreadsheets, CSV files, log files and databases, are widely used to represent and share information. However, tables remain beyond the scope of regular text processing systems which often treat them like free text.

This dissertation presents TABEL — a domain independent and extensible framework to infer the semantics of tables and represent them as RDF Linked Data. TABEL captures the intended meaning of a table by mapping header cells to classes, data cell values to existing entities and pair of columns to relations from an given ontology and knowledge base. The core of the framework consists of a module that represents a table as a graphical model to jointly infer the semantics of headers, data cells and relation between headers. We also introduce a novel Semantic Message Passing scheme, which incorporates semantics into message passing, to perform joint inference over the probabilistic graphical model. We also develop and explore a “human-in-the-loop” paradigm, presenting plausible models of user interaction with our framework and its impact on the quality of inferred semantics.

We present techniques that are both extensible and domain agnostic. Our framework supports the addition of preprocessing modules without affecting existing ones, making TABEL extensible. It also allows background knowledge bases to be adapted and changed based on the domains of the tables, thus making it domain independent. We demonstrate the extensibility and domain independence of our techniques by developing an application of TABEL in the healthcare domain. We develop a proof of concept for an application to generate meta-analysis reports automatically, which is built on top of the semantics inferred from tables found in medical literature.

A thorough evaluation with experiments over dataset of tables from the Web and medical research reports presents promising results.

Committee: Drs. Tim Finin (chair), Tim Oates, Anupam Joshi, Yun Peng, Indrajit Bhattacharya (IBM Research) and L. V. Subramaniam (IBM Research)

Photo by Marlayna Demond. Sitting (left to right): Joel DeWyer, Business Manager; Dr. Alan T. Sherman, Director and GM Sam Palatnik, Coach. Standing (left to right): GM Akshayraj “The Indian Knight” Kore; IM Levan “The Georgian Gangster” Bregadze; GM Niclas “The Dark Knight” Huschenbeth, Captain; IM and WGM Nazi “The Black Widow” Paikidze and IM Tanguy “The Belgium Butcher” Ringoir.

The UMBC chess team won its match (4-0) against Lindenwood University in the first round of the 2014 Pan American Intercollegiate Team Chess Championship being held this week in South Padre Island, Texas. Five more rounds will be played over the next three days to determine winner and the top four USA schools, which will compete in the President’s Cup championship April 5-6 at the New York Athletic Club to determine the U.S. college team chess champion.

UMBC’s chess team has won or tied for first place a record ten times at the Pan-Am and a record six times at the President’s Cup. The UMBC team in this year’s Pan-Am consists of GM Niclas Huschenbeth, IM Levan Bregadze, IM Tanguy Ringoir, GM Akshayraj Kore and WGM Nazi Paikidze (alternate).

CSEE professor Alan Sherman, director of UMBC’s chess program, reports that the field is especially strong this year with 22 International Grandmasters (GM) or International Woman Grandmaster (WGM) and notes that “At least one strong school will not make the Final Four”.

You can view the latest results on the tournament website and watch selected games live on Monroi.

Update 12/28: UMBC beat Toronto 3-1 in round two on Sunday morning and plays Texas Tech – C in round three on Sunday evening.

Our Health IT students could be highly qualified for this due to their strong technical skills.
It can be a valuable experience.

Dr. Koru

Position Summary:
    The Clinical Informatics Program at the Mayo Clinic Kern Center for the Science of Health Care Delivery (http://www.mayo.edu/research/centers-programs/robert-d-patricia-e-kern-center-science-health-care-delivery) is offering a 3-6 month Summer/Fall internship in clinical informatics for 2015.  Several projects are currently under development in the areas of data integration, information retrieval and knowledge management using electronic health records, big data analysis, prediction modeling, and clinical decision support. We are seeking candidates with computer science/informatics background and experience in data mining, machine learning, and big data.  Research activities and successful findings will be translated into development activities for existing projects, peer-reviewed publications, and extramurally funded research programs.
    
    Basic Requirements:
    - Enrollment or recently completed a MS/PhD in Computer Science, Mathematics/Statistics, Biomedical Informatics, Information Technology, or related field
    - Research experience in one or more of the following areas: Data mining and machine learning, Electronic health records, Big data technologies (Apache Hadoop, Pig, Storm)
    - Coding skills in Java/C/C++ and knowledge of a scripting language (Perl, Python, JavaScript)
    
    Specific skills in the following environments and technologies are desirable:
    - R, SAS, Matlab
    - Oracle, MySQL, MongoDB, Virtuoso
    - Data mining toolkits (Weka, Mallet)
    The candidate should be highly motivated, willing to learn and demonstrate initiative in assigned tasks. Must possess excellent analytical, problem-solving, and technical design skills. Strong oral and written communication and interpersonal skills are required.
    
    How to apply:
    To apply, send cover letter and resume to:
    Che Ngufor, PhD
    Post-doctoral Research Associate
    Division of Biomedical Statistics and Informatics
    Department of Health Sciences Research
    Mayo Clinic
    200 First Street SW, Rochester, MN  55905, USA
    Email: Ngufor.Che@mayo.edu<mailto:Ngufor.Che@mayo.edu>
    
    Further information:
    For additional information about the position, contact Jyotishman Pathak, PhD (Pathak.Jyotishman@mayo.edu<mailto:Pathak.Jyotishman@mayo.edu>) or Che Ngufor, PhD (Ngufor.Che@mayo.edu<mailto:Ngufor.Che@mayo.edu>). The application deadline for this internship is open until filled.
    
    
    Jyotishman Pathak, PhD
    Associate Professor of Biomedical Informatics
    Director, Clinical Informatics, Kern Center for the Science of Health Care Delivery
    Director, Biomedical Informatics, Center for Clinical and Translational Science
    Department of Health Sciences Research
    Mayo Clinic
    200 1st Street SW, Rochester, MN 55905, USA
    Phone: +1-507-538-8384
    Fax: +1-507-284-0360
    Email: Pathak.Jyotishman@mayo.edu<mailto:Pathak.Jyotishman@mayo.edu>
    Homepage: http://jyotishman.info<http://jyotishman.info/>