There have been many models developed for IR: Boolean model, vector space model, probabilistic model and more recent statistical language model. This lecture will provide a short overview of these models. More emphasis will be put on recent work trying to further improve the existing models in order to account for the language problems in search. In particular, we will review studies in two directions, which try to integrate relationships between different terms and dependencies between terms within the same document or query. We will end the lecture with a discussion on the remaining problems in IR models.
Jian-Yun Nie is a professor in University of Montreal. He has been working in the area of IR and NLP for more than 25 years. He has worked on a range of topics such as IR models, cross-language IR, IR using query logs, and so on, and has published more than 150 papers in journals and conferences. He is on board of 7 journals in the areas of IR and NLP, and is a PC member of many conferences. He is a general chair of the SIGIR 2011 conference in Beijing. More information can be found at http://www.iro.umontreal.ca/~nie/