On May 31, three teams from USC were presented with first and third place awards in the 2012 AI Mashup Challenge, part of the annual Extended Semantic Web Conference. The team’s entries were course projects for this spring’s CSCI 548 “Information Integration on the Web,” taught by Research Assistant Professors Pedro Szekely, Jose-Luis Ambite and Research Professor Craig Knoblock, who encouraged students to submit their work to the Challenge.
Voting results are tallied for the 2012 AI Mashup Challenge.
The AI Mashups Challenge, held in Crete this year, recognizes participants’ ingenious use of artificial intelligence technology in lightweight web applications offering new functionality through the combination, aggregation and transformation of resources and services available on the web (or “mashups”). Other applications of AI mashup technology include smart systems that do simple household chores, search engines that combine general dictionaries with specialized jargon to produce better search results and the display of environmental articles on a map according to relevant geography.
For their efforts, the team behind first place mashup CareerGalaxy went home with an award and prize of € 1,750 (about 2,143 USD), while the teams behind mashups Hugos and WhatsGoingOn, which tied for third place, were awarded ten O’Reilly e-books and ten Addison-Wesley computer science books respectively.
USC Viterbi student Yiyu Li accepts first prize for her team's mashup app, CareerGalaxy.
CareerGalaxy, an online planner that charts potential career paths for users, was developed by USC Viterbi students Yiyu Li, Xilong Fan, Mamta Bhavsar and Yemin Shi. The application takes a user’s college major and degree and creates colorful flowcharts providing details of potential career paths. It does this by mining the data available from professional networking site LinkedIn, which has a huge amount of data on the degrees and career paths people have followed. Yiyu Li attended and presented the winning project at the conference.
Created by students Yong Jin, Yupeng Ji and Shiliang Xuea, Hugos addresses the difficulty of displaying large numbers of real estate listings on a browser’s limited display area. Adapting maps of listings according to user preference and a smart recommendation system, Hugos dynamically adjusts the display based on user interests and predicts missing listing values according to house price and size.
WhatsGoingOn, produced by students Wenjing Yu, Xuming Zhu and Jose Romero-Lobo, examines and displays trends between news headlines that appear on Twitter and those of major media outlets and searches for correspondences between the two.
