Tetherless World Constellation

Two members of the Rensselaer Polytechnic Institute science faculty have been selected as fellows of the American Association for the Advancement of Science (AAAS). Susan Gilbert, professor and head of biology, and James Hendler, senior constellation professor in the Tetherless World Constellation and head of the information technology and web science program, are two of the 539 newly selected AAAS fellows. They were recognized for their efforts to advance science or its applications that are deemed scientifically or socially distinguished, according to AAAS. The announcement will be made in the Dec. 23, 2011, issue of the journal Science.

Elisa Kendall, Partner – Thematix LLC and Deborah McGuinness, Tetherless World Chaired Constellation Professor – Rensselaer Polytechnic Institute (RPI) will be giving the featured presentation at SemTechBiz DC The conferences takes place at the Kellogg Conference Hotel in Washington, DC November 29-December 1, 2011, and will feature highly-anticipated sessions led by top thinkers in the fields of Open Government, Content Management, Enterprise Data Management, Linked Data, and much more. Deborah L. McGuinness

A talk by Shinichiro Nakamura, Faculty of Political Science and Economics, Waseda University, Tokyo, Japan and Ecotopia Science Institute, Nagoya University, Nagoya, Japan. The talk will take place at 4:00pm on Wednesday, November 16, 2011 in room 1140 of the Winslow Building, Rensselaer Polytechnic Institute. Read more...

A talk by Boyan Brodaric, Research Scientist, Geological Survey of Canada in Ottawa Canada. The talk will take place at 1:30pm on Tuesday, November 15, 2011 in room 1140 of the Winslow Building, Rensselaer Polytechnic Institute. Read more...

A talk by Vasco Furtado, Professor, Computer Science at the University of Fortaleza, Brazil. The talk will take place at 10:30am on Tuesday, November 8, 2011 in room 1140 of the Winslow Building, Rensselaer Polytechnic Institute. Read more...

The annual I-SEMANTICS Conference hosted its fourth Triplification Challenge, an event aimed at stimulating the availability of large quantities of RDF data and showcasing practical applications built on that data. The Challenge consisted of an unspecific open data track and a dedicated open government data track for which winners were selected. The prize money of 1000 Euro each was sponsored by Wolters Kluwer Germany. The open government data track award was won by the TWC team of John Erickson, Yongmei Shi, Li Ding, Eric Rozell, Jin Zheng and Jim Hendler for their contribution TWC International Open Government Dataset Catalog (IOGDC). Ph.D. student Eric Rozell presented the team's work and accepted the award on behalf of the team. Eric's innovative S2S faceted browser is featured in the IOGDC live demo.

Medha Atre is now Dr. Atre after passing her thesis defense on August 19th. She is the first Tetherless World Constellation doctoral student to receive a Ph.D. since the constellation was formed 4 years ago by Dr. Jackson.

Tetherless World Constellation
Department of Computer Science
Thesis Advisor: Professor James A. Hendler

Bit-by-Bit: Indexing and Querying RDF data using Compressed Bit-Vectors

Friday, August 19, 2011
Winslow 1140 - 2:30 p.m.

The Resource Description Framework (RDF) is widely being adapted as a for information representation in various domains, e.g., biotechnology (UniProt), government data (the data.gov project), scholarly resources (DBLP), web resources (DBPedia), connection between people (FOAF) and many more. Due to the increasing use of RDF as a standard for data representation in the past few years, the amount of RDF data available on the web has increased at a break-neck speed. This has necessitated addressing two important issues -- efficient ways of (a) storing, and (b) querying the RDF data.

In this thesis, we propose a novel way of storing RDF data and an efficient way of processing SPARQL join queries -- BitMat -- which is specifically aimed at low-selectivity SPARQL join queries. BitMat uses the well-known technique of compressed bit-vectors to store RDF data by representing it as a 3-dimensional bit-cube (subject, predicate, object as each dimension of the bitcube). The key aspect of our join query processing is a novel 2-phase algorithm combining the idea of semi-joins and multi-way joins.

We have also extended BitMat's technique of /pruning/ the candidate RDF triples to process DISTINCT and OPTIONAL clauses in SPARQL. Specifically for the DISTINCT clause, BitMat's pruning phase can be used to omit BitMats that are not needed for final result generation. This further enhances the memory requirements of the query processor.

We have also given a detailed description of how other SPARQL clauses can be handled -- either using BitMat algorithm, or by working on the results of BitMat algorithm.

RDF data is primarily graph data, hence we have also proposed to extend the technique of compressed bit-vectors used in BitMat to solve /label order constrained reachability/ queries on RDF graph. We propose a system -- BitPath -- which uses compressed bit-vectors to build specific indices for each node in the RDF graph and a query processing algorithm based on greedy pruning and divide-and-conquer approach to solve this problem.

Thus in this thesis we are addressing two important aspects of querying RDF data -- (a) performance intensive SPARQL queries, and (b) label-order-constrained-reachability queries. We show that using characteristics of compressed bit-vectors in both -- BitMat and BitPath algorithms -- we can come up with efficient solutions to both the problems.

I’M not the most talkative person on airplanes. I used to be much friendlier, but an inevitable question is, “What do you do?” When I say, “I’m a professor of computer science,” I always wind up doing a lot of free consulting.

Right now the race is on to develop search that is both more accurate and more customized. What if there was an app that could act as an agent for you, taking all the meta-data from searchable pages, filtering and arranging them to make sense for you, and then giving you the perfect search result? You and your computer would work cooperatively to find exactly what you were looking for. This idea is called the semantic web – the ability to do search with meaning, rather than just syntax. Deborah McGuinness is a Professor of Computer Science and Cognitive Science at Rensselaer Polytechnic Institute, and the Chair of the Tetherless World Constellation. She tells Nora about a wine agent she’s developed as an early semantic web demonstration.

Create apps; Win Prizes!

The Tetherless World Constellation at RPI is excited to announce that TWC and Elsevier's SciVerse Developer Network will be holding a 24-hour Health and Life Sciences Semantic Web Hackathon 27-28 June 2011. The Elsevier-sponsored event will be held at the beautiful Pat's Barn on the campus of the Rensselaer Technology Park in Troy, NY.

Participants will compete with each other to develop apps using linked data from TWC and other sources, web APIs from Elsevier SciVerse, and visualization and other resources from around the Web.

Event website http://tw.rpi.edu/web/event/TWCElsevierHackathonJune2011 Registration at: http://twcsciverse2011.eventbrite.com/

Titled "Fundamental Analysis Powered by Semantic Web", authors Xian Li, Jie Bao, and James Hendler won best paper at the IEEE Symposium on Computational Intelligence for Financial Engineering & Economics 2011, held in Paris, France.

Despite the complexities associated with semantic technologies, efforts to adopt the approach for drug development are bearing fruit, according to several presentations at last week's Conference on Semantics in Healthcare and Life Sciences in Cambridge, Mass.

This year's conference began with a series of hands-on tutorials coordinated by Joanne Luciano, a research associate professor at Rensselaer Polytechnic Institute, that were intended to show how the technology can be used to address drug development needs.

During the tutorials, participants used semantic web tools to create mashups using data from the Linked Open Data cloud and semantic data that they created from raw datasets. Participants were shown how to load data into the subject-predicate-object data structure dubbed the "triple store;" query it using the semantic query language SPARQL; use inference to expand experimental knowledge; and build dynamic visualizations from their results.

RPI Tetherless World Senior Constellation Professor Dr. Deborah McGuinness discusses ongoing research in the development of smarter phones.