Finally I find some time to write the trip report for SIGIR 2009 and wish to share some of my experiences with you. This year's
SIGIR was held in Boston, MA, USA from July 19 to July 23, 2009.
Here I list some nice talks, presentations and interesting topics I found during the conference.
1. Large scale distributed systems for Information Retrieval
It emphasizes the scalability and the tradeoffs associated with distributed processing on large datasets. You can have a closer look at the following talks or papers:
1.1 Tutorial: Data-Intensive Text Processing with MapReduce
1.2 Workshop: Large-Scale Distributed Systems for Information Retrieval
1.3 Research papers
1.3.1 Quantifying Performance and Quality Gains in Distributed Web Search Engines (Federated, distributed search session)
1.3.2 Brute Force and Indexed Approaches to Pairwise Document Similarity Comparisons with MapReduce (Efficiency session)
2. Ad search
Advertisements have become the main revenue for search engine companies. Ad search has become an important research aspect in the IR community. If you have interests, you can refer to:
2.1 Workshop: Information Retrieval and Advertising
2.2 Industry Talk: Ad Retrieval – A New Frontier of Information Retrieval by Vanja Josifovski, Yahoo! Research
2.3 Research papers
2.3.1 Efficient Query Expansion for Advertisement Search (Expansion and feedback session)
2.3.2 Optimizing Search Engine Revenue in Sponsored Search (Web Retrieval session)
3. Vertical search
Besides Ad search, vertical search has become another new trend in the IR community. Vertical search aims at providing high accurate results for a specific domain, a specific type, or inside a specific organization. You can refer to:
3.1 Panel on vertical search: What concerns drive the business of enterprise search, and how should technologies approach them?
What are the technical challenges driving enterprise search, and what are/should researchers and practitioners be doing to address them?
3.2 Research papers
3.2.1 Sources of Evidence for Vertical Selection (Best paper, Vertical search session)
3.2.2 Adaptation of Offline Vertical Selection Predictions in the Presence of User Feedback (Vertical search session)
3.2.3 in the vertical search session, , , in the federated, distributed search, Effective Query Expansion for Federated Search
4. User interaction and query understanding
The nature of the search is to understand the user information needs, understand the content or description of any documents or entities, and try to match the needs and contents accordingly. You can pay attention to the following talks or presentations:
4.1 Tutorial: Web Query Log Mining
4.2 Workshop: Understanding the user - Logging and interpreting user interactions in information search and retrieval
4.3 Research papers
4.3.1 Predicting User Interests from Contextual Information (Interactive search session)
4.3.2 A Comparison of Query and Term Suggestion Features for Interactive Searching (Interactive search session)
4.3.3 An Aspectual Interface for Supporting Complex Search Tasks (Interactive search session)
4.3.4 Segment-Level Display Time as Implicit Feedback: A Comparison to Eye Tracking (Expansion and feedback session)
4.3.5 Query Dependent Pseudo-Relevance Feedback based on Wikipedia (Expansion and feedback session)
5. Others
There are other interesting topics like social media, Web 2.0, question answering and multimedai. If you have interests, you can refer to the homepage of SIGIR 2009 http://www.sigir2009.org for more details. Here, I would like say something about two industry talks:
One is "Corpus Linguistics and Semantic Technology at the New York Times" and the other is "Developments in Bing". Both of them leveraged structured data or even linked data in form of RDF for search improvements. I think this is good news for the Semantic Web community, especially for those who have intersts in semantic search.
[返回] W3CHINA.ORG讨论区 - 语义网·描述逻辑·本体·RDF·OWL →
W3CHINA.ORG讨论区 - Web新技术讨论 → 『 Semantic Web(语义Web)/描述逻辑/本体 』 → Trip report for SIGIR 2009
(订阅本版)
2009/8/26 11:37:00
新浪微博
Trip report for SIGIR 2009
