CS6200 Project Three

Assigned:

Friday, April 4, 2014

Due:

10:00pm, Friday, April 18, 2014

important Submit by committing your files to svn. See the checklist below to make sure you turn everything in.

Overview

In this project, you will read two research papers on a common theme and write a report which both summarizes the papers and presents your thoughts for extending the work.

Report

Write a report based on the papers in one of the following categories. Your report should be written for a professional audience which is familiar with Computer Science research but has not studied these particular papers. The average report length should be around 5 pages. Feel free to include particularly interesting figures from the papers, or of your own, if you wish.

important Please submit your report as a PDF.

1. Begin with an introductory section which gives a very high-level summary of the two papers and their relationship. If you wish, you can provide further context or background for the papers here.
2. Explain each of the two papers in its own section. What was the main contribution of this paper? What methods do the authors propose? How was it evaluated, and did the authors present a convincing case? If you choose a survey paper you don't have to understand or explain all the methods discussed. Instead, compare at least three selected methods.
3. Explain the shortcomings of each paper. What work remains to be done?
4. Present some ideas for how to extend the work. If you are presenting the authors' own suggestions for future work that's OK, but you must contribute some novel thoughts of your own to the suggestion.
5. Conclude with a summary that expresses your personal opinion on the papers in a few sentences.

Option One: Ad Placement

Using auctions to select ads to display on a search engine. These papers are from Hal Varian, the chief economist at Google and a founding dean of the School of Information at U.C. Berkeley.

• Position Auctions. Hal R. Varian, 2007.
• Online Ad Auctions. Hal R. Varian, 2009.

Option Two: Spam Detection

Identifying spam is a key aspect of improving search quality.

• Survey on web spam detection: principles and algorithms. Nikita Spirin, Jiawei Han, 2012.
• Don't Follow Me: Spam Detection in Twitter. Alex Hai Wang, 2010.

Option Three: Knowledge Building

Parsing online text to automatically generate a large database of facts. These papers are from the Never-Ending Language Learning project at Carnegie Mellon.

• Toward an Architecture for Never-Ending Language Learning. Andrew Carlson, Justin Betteridge, Bryan Kisiel, Burr Settles, Estevam R. Hruschka Jr., and Tom M. Mitchell, 2010.
• Random Walk Inference and Learning in A Large Scale Knowledge Base. Ni Lao, Tom M. Mitchell, William W. Cohen, 2011.

Option Four: Question Answering

Question Answering is a search task in which you present the user with the answer to a question, instead of a list of documents.

• Learning surface text patterns for a Question Answering system. Deepak Ravichandran, Eduard Hovy, 2002.
• Is Question Answering fit for the Semantic Web?: A survey. Vanessa Lopez, Victoria Uren, Marta Sabou, Enrico Motta, 2011.

Option Five: Learning to Rank

These papers focus on LambdaMART, one of the current most popular LtR methods.

• From RankNet to LambdaRank to LambdaMART: An Overview. Christopher J.C. Burges, 2010.
• Query-biased learning to rank for real-time twitter search. Xin Zhang, Ben He, Tiejian Luo, Baobin Li, 2012.