Entries tagged with “papers” from Computer Science Department News

Computer Science Department doctoral student Beibei Yang presented a research paper co-authored with Prof. Jesse Heines, Domain-Specific Semantic Relatedness from Wikipedia: Can a Course be Transferred?, at the North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL-HLT) 2012 conference. The event was held June 3–8, 2012 in Montréal (Québec), Canada.

The conference covered a broad spectrum of disciplines working towards enabling intelligent systems to interact with humans using natural language, and towards enhancing human-human communication through services such as speech recognition, automatic translation, information retrieval, text summarization, and information extraction.

Yang and Heines analyzed the problem of transferring credits across undergraduate institutional. About 1/3 of all college students in the U.S. transfer between institutions. In their work, Yang and Heines proposed a Wikipedia-based domain-specific semantic relatedness measure that analyzes course descriptions to suggest whether a course can be transferred from one institution to another.

They showed that the proposed work received a high correlation of 0.85 when compared to human judgment on computer science courses. And it only took less than 1 minute to compare one pair of courses on a standard laptop system.

Their poster at the conference attracted many researchers from universities and organizations including CMU, Stanford, University of Edinburgh, Google, IBM research, and Nuance.

Yang also received a travel grant of $500 from the conference.

yang-naacl-hlt.jpg
Ph.D. student Beibei Yang at the NAACL-HLT 2012 student research workshop. (Courtesy Andy Dufilie)
A research paper by Computer Science Department doctoral student Harshavardhan Achrekar and co-authors Avinash Gandhe, Ross Lazarus, Ssu-Hsin Yu, and Benyuan Liu, Twitter improves Seasonal Influenza Prediction, has been selected as the Best Student Paper at the Fifth International Conference on Health Informatics (HEALTHINF 2012) held February 1 to 4, 2012 in Vilamoura, Algarve, Portugal. Fewer than 9 percent of submitted manuscripts were accepted as full papers.

The purpose of the International Conference on Health Informatics was to bring together researchers and practitioners interested in the application of information and communication technologies (ICT) to healthcare and medicine in general and to the specialized support to persons with special needs in particular. Databases, networking, graphical interfaces, intelligent decision support systems and specialized programming languages were just a few of the technologies currently used in medical informatics.

Achrekar is utilizing information posted on Online Social Networks (OSNs) such as Twitter and Facebook to help improve the prediction of influenza levels within United States population and keep track of its spread. He has designed and implemented a framework called the Social Network-Enabled Flu Trends (SNEFT), which is used to continuously monitor flu-related messages, extract relevant location and user demographic information, track and predict the flu conditions in real time.

Since 2009, Achrekar has tapped into Twitter and extracted millions of influenza-related user posts to date, providing an almost-instantaneous snapshot of current epidemic conditions. Using comprehensive mathematical models, the framework can estimate nationwide as well as regional and age based flu activity with high accuracy.

This research is supported by a $200,000 grant from the National Institutes of Health under a Small Business Innovation Research Award. Results presented in this scientific publication show that these posts on Twitter closely match the number of flu-like cases reported by the Center for Disease Control and Prevention (CDC), and can significantly enhance public health preparedness against influenza and other large-scale pandemics.

harsh_award.jpg
Harsh Achrekar (left) and faculty adviser Prof. Benyuan Liu showcasing HEALTHINF 2012 best paper award.
Associate Professor Karen Daniels and her doctoral student Shu Ye presented their paper “Hierarchical Delaunay Triangulation for Meshing” at the 10th International Symposium on Experimental Algorithms in May 2011 in Greece.

Their paper discusses an elliptical pad structure and its polygonal approximation. The elliptical pad is a part of via model structures, which are critical components on today’s multilayered Printed Circuit Board (PCB) and electrical packaging.

To explore meshing characterization of the elliptical pad helps mesh generation over 3D structures for electromagnetic modeling and simulation on PCB and electrical packaging. Because elliptical structures are often key PCB features, the authors introduce a hierarchical mesh construct and show that it has several useful quality characteristics related to Delaunay triangulation.

The Delaunay triangulation has the important modeling characteristic that the minimum triangle angle is maximized.

Daniels and Ye then show experimentally that the Computational Geometry Algorithm Library’s meshing of an elliptical structure at different resolution levels and with various aspect ratios produces patterns similar to the hierarchical construct. In particular, the experiment also shows that the result of meshing is not only a constrained Delaunay triangulation, which preserves the segments that approximate the elliptical structure, but, significantly, it is also a Delaunay triangulation.

A copy of the paper may be downloaded here.

SEA2011_Blog_Image.png

Examples of hierarchical Delaunay triangulations.

Doctoral candidate Kate Tsui, a member of Prof. Holly Yanco's Robotics Lab, presented research on telepresence robots at the 6th Annual ACM/IEEE International Conference on Human-Robot Interaction in Lausanne, Switzerland in March 2011. The conference was highly selective, with an acceptance rate of 22%.

The paper, Exploring Use Cases for Telepresence Robots, was co-authored by Kate Tsui, Munjal Desai, Holly Yanco, and Chris Uhlik of Google. The studies of the telepresence robots were conducted by Kate and Munjal at Google in Mountain View, California, in July and August 2010.

hugo_lightuptie.jpg
Tsui’s dissertation research focuses on designing telepresence robot systems for use by people with special needs. In continuing work, the VGo robot, above, has been augmented with additional sensors and processing using a BeagleBoard. VGo’s custom necktie, designed by Adam Norton, lights up and provides status to the user.
Alexander Baumann, a PhD candidate advised by Prof. Georges Grinstein, will present his paper, “Enhancing STEM classes using Weave: a collaborative web-based visualization environment,” at ISEC (Integrated STEM Education Conference), held in Ewing, NJ, April 2nd.

ISEC is co-located with the Trenton Computer Festival, and is intended to encourage "thought-provoking discussions on how the 'silos' separating instruction in and, ultimately, comprehension among STEM fields can be removed and, subsequently, on the various ways 'integrated STEM methods' take shape in K-16 classrooms."

Baumann described how Weave (WEb-based Analysis and Visualization Environment) could be used to support STEM education by way of collaborative visual analysis.  

Weave is a framework that provides advanced data visualization capabilities on the web. Some unique features of the system are high flexibility in aesthetic and interaction choices in the software to handle a wide range of audiences; session histories to record user actions taken in the system for use in restoring, reviewing, and analyzing previous work; and collaboration capabilities to allow multiple users at the same or different locations to work together from separate computers.  

The use of collaborative tools to support STEM learning is a major focus of this paper. Students can learn from and interact with the teacher in live training sessions, break off into groups to study data sets and provide results, and use smaller subgroups to divide up work that can be shared with the group at any time.  


baumann-weave.png
A screenshot of data loaded in Weave (with some modifications for presentation purposes).
UMass Lowell's robotics faculty and students will be presenting four new papers at the IEEE International Conference on Technologies for Practical Robot Applications (TePRA). The conference will be held on April 11 and 12, 2011 in Woburn, MA.

The TePRA conference is based in the greater Boston area, and has the goal of introducing students to “the state-of-the-art in practical robotics R&D and representatives of both industry and the military.”

Prof. Holly Yanco's Robotics Lab will be presenting two papers: Hand and Finger Registration for Multi-Touch Joysticks on Software-Based Operator Control Units, by Mark Micire, Eric McCann, Munjal Desai, Katherine M. Tsui, Adam Norton, and Holly A. Yanco, and Essential Features of Telepresence Robots, by Munjal Desai, Katherine M. Tsui, Holly A. Yanco, and Chris Uhlik.

Prof. Yanco has a third paper based on her collaborations at MITRE, Using Geographic Information Systems (GIS) for UAV Landings and UGV Navigation, by Nathan Rackliffe, Holly A. Yanco, and Jennifer Casper.

Prof. Fred Martin's doctoral student Haiyang Zhang will present the fourth paper, entitled Robotic Mapping Assisted by Local Magnetic Field Anomalies

Prof. Yanco notes that when a previous robotics student, Erin Rapacki, presented at TePRA two years ago, “her work was picked up by New Scientist and Popular Science based on exposure at this conference.”

So it promises to be a high visibility event, with excellent representation this year from UMass Lowell.
Derfen Lin, a PhD candidate advised by Prof. Jie Wang, presented her paper, “Producing Automated Mosaic Art Images of High Quality with Restricted and Limited Color Palettes,” at the IMAGAPP conference, held in Algarve, Portugal, March 5-7, 2011.

IMAGAPP is part of VISIGRAPP, a joint international conference on computer vision, imaging, and computer graphics theory and applications. 

Lin studied the problem of mosaic art images, which are made from bricks, tiles, or counted cross-stitch patterns.

Artists need to divide the original image into small parts of reasonable sizes and shapes, and represent the colors of each part using just one closest color selected from a given color palette, a process called “dithering.”

Using standard methods to automate this process, the resulting mosaic image may contain undesirable visual artifacts of patches and color bandings.

In her study, Lin presented a new error-diffusion scheme, called Four-Way Block dithering (FWB), which corrects certain artifacts caused by existing dithering methods.

Lin showed that FWB can better retain the original structure and reduce unstructured artifacts.  She also showed that FWB dithering produces much better peak signal-to-noise ratios on mosaic images over those generated by previous methods (please see image below for an example).


fwb.png
Image (a) shows dithering with previous approach; it is clearly noticeable that Mona Lisa’s right eye is blurry and looks half-closed. Image (b) shows use of the FWB algorithm. The eye problem is corrected. 
Bo Yan, a doctoral student advised by Prof. Guanling Chen, will present his paper on AppJoy at the Ninth ACM International Conference on Mobile Systems, Applications, and Services (MobiSys). MobiSys is a prestigious conference on mobile and wireless systems with a low acceptance rate of 18% this year.

AppJoy is a service that helps Android users to discover interesting applications through personalized recommendation algorithms. It is available in Åndroid Market and has been used by more than 5000 users worldwide. For more information, see the AppJoy web site.

Subscribe to feed Subscribe to this blog's feed

Subscribe to feed Search results matching “papers”