Information Retrieval involves finding relevant information for user queries, ranging from simple domain of database search to complicated aspects of web search (Eg - Google, Bing, Yahoo). Currently, researchers are developing algorithms to address Information Need of user(s), by maximizing User and Topic Relevance of retrieved results, while minimizing Information Overload and retrieval time.

• Introduction to Modern Information Retrieval - G.G. Chowdhury. Neal-Schuman, 2003. (Intended for students of library and information studies).

• Indri Search Engine - Another Open Source Search Engine competitor of Apache Lucene.
  • Lemur Toolkit - Open Source Toolkit for research in Language Modeling, filtering and categorization.

Standard IR Collections

• Blog - Explore information seeking behavior in the blogosphere.
• Chemical IR - Address challenges in building large chemical testbeds for chemical IR.
• Clinical Decision Support - Investigate techniques to link medical cases to information relevant for patient care.
• Confusion - Study Known Item Searching problem.
• Contextual Suggestion - Investigate search techniques for complex information needs (context and user interests based).
• Crowdsourcing - Explore crowdsourcing methods for performing and evaluating search.
• Enterprise - Study search over the organization data.
• Entity - Perform entity-related search (find entities and their properties) on Web data.
• Filtering - Binarily decide retrieval of new incoming documents given a stable information need.
• Federated Web Search - Study merge performance for results from various search services.
• Genomics - Study retrieval efficiency of genomics data and corresponding documentation.
• HARD - Obtain High Accuracy Retrieval from Documents by leveraging searcher's context.
• Interactive Track - Study user interaction with text retrieval systems.
• Knowledge base acceleration - Study algorithms that improve efficiency of human Knowledge Base.
• Legal Track - Study retrieval systems that have high recall for legal documents use case.
• Medical Track - Explore unstructured search performance over patients record data.
• Microblog Track - Examine satisfaction of real-time information need for microblogging sites.
• Million Query Track - Explore ad-hoc retrieval over large set of queries.
• Novelty Track - Investigate systems' abilities to locate new (non-redundant) information.
• Question Answering Track - Test systems that scale beyond document retrieval, to retrieve answers to factoid, list and definition type questions.
• Relevance Feedback Track - For deep evaluation of relevance feedback processes.
• Robust Track - Study individual topic's effectiveness.
• Session Track - Develop methods for measuring multiple-query sessions where information needs drift.
• SPAM Track - Benchmark spam filtering approaches.
• Tasks Track - Test if systems can induce possible tasks, users might be trying to accomplish for the query.
• Temporal Summarization Track - Develop systems that allow users to efficiently monitor the information associated with an event over time.
• Terabyte Track - Test scalability of IR systems to large scale collection.
• Web Track - Explore information seeking behaviors common in general web search.
• CLIR Test Collections - This dataset can be used for cross lingual IR between CJKE (Chinese-Japanese-Korean-English) languages. It is suitable for the following tasks:
  • Multilingual CLIR
  • Bilingual CLIR
  • Single Language CLIR
• Cross Language Q&A (CLQA) dataset collection - It supports following bi-lingua and mono-lingua:
  • Bi-lingua
    • Japanese to English.
    • Chinese to English.
    • English to Japanese.
    • English to Chinese.
  • Mono-lingua
    • Chinese to Chinese.
    • Japanese to Japanese.
    • English to English.
• Advanced Cross Linugal Information Retrieval and Question Answering (ACLIA) - The dataset is used for the task of cross-lingual question answering but the complexity of the task is higher than CLQA dataset.
• AdHoc - News Test suite.
• Domain Specific Test Suite - On collections of scientific articles.
• Question Answering Test Suite.
• RCV1 (Reuter's Corpus Volume 1) - Consists of only English language News stories.
• RCV2 (Reuter's Corpus Volume 2) - Consists of stories in 13 languages (Dutch, French, German, Chinese, Japanese, Russian, Portuguese, Spanish, Latin American Spanish, Italian, Danish, Norwegian, and Swedish). Note that the stories are not parallel.
• TRC (Thomson Reuters Text Research Collection) - This is a fairly recent corpus consisting of 1,800,370 news stories covering the period from 2008-01-01 00:00:03 to 2009-02-28 23:54:14.

External Curation Links

Technical Talks

Philosophical Talks

• Web Search and Data Mining Conference - WSDM.
• Special Interests Group on Information Retrieval - SIGIR.
• Text REtrieval Conference - TREC.
• European Conference on Information Retrieval - ECIR.
• World Wide Web Conference - WWW.
• Conference on Information and Knowledge Management - CIKM.
• Forum for Information Retrieval Evaluation - FIRE.
• Conference and Labs of the Evaluation Forum - CLEF.
• NII Testsbeds and Community for Information access Research - NTCIR.

Interesting Reads