Session Based Click Features for Recency Ranking.

User behavior driven ranking without editorial judgments.

Online learning for recency search ranking using real%2Dtime user feedback.

Towards recency ranking in web search.
In web search, recency ranking refers to ranking documents by relevance which takes freshness into account. In this paper, we propose a retrieval system which automatically detects and responds to recency sensitive queries. The system detects recency sensitive queries using a high precision classifier. The system responds to recency sensitive queries by using a machine learned ranking model trained for such queries. We use multiple recency features to provide temporal evidence which effectively represents document recency. Furthermore, we propose several training methodologies important for training recency sensitive rankers. Finally, we develop new evaluation metrics for recency sensitive queries. Our experiments demonstrate the efficacy of the proposed approaches.
A model to estimate intrinsic document relevance from the clickthrough logs of a web search engine.

Early exit optimizations for additive machine learned ranking systems.
Some commercial web search engines rely on sophisticated machine learning systems for ranking web documents. Due to very large collection sizes and tight constraints on query response times, online efficiency of these learning systems forms a bottleneck. An important problem in such systems is to speedup the ranking process without sacrificing much from the quality of results. In this paper, we propose optimization strategies that allow short-circuiting score computations in additive learning systems. The strategies are evaluated over a state-of-the-art machine learning system and a large, real-life query log, obtained from Yahoo!. By the proposed strategies, we are able to speedup the score computations by more than four times with almost no loss in result quality.
Empirical Exploitation of Click Data for Task Specific Ranking.

Global ranking by exploiting user clicks.
It is now widely recognized that user interactions with search results can provide substantial relevance information on the documents displayed in the search results. In this paper, we focus on extracting relevance information from one source of user interactions, i.e., user click data, which records the sequence of documents being clicked and not clicked in the result set during a user search session. We formulate the problem as a global ranking problem, emphasizing the importance of the sequential nature of user clicks, with the goal to predict the relevance labels of all the documents in a search session. This is distinct from conventional learning to rank methods that usually design a ranking model defined on a single document; in contrast, in our model the relational information among the documents as manifested by an aggregation of user clicks is exploited to rank all the documents jointly. In particular, we adapt several sequential supervised learning algorithms, including the conditional random field (CRF), the sliding window method and the recurrent sliding window method, to the global ranking problem. Experiments on the click data collected from a commercial search engine demonstrate that our methods can outperform the baseline models for search results re-ranking.
Enhancing topical ranking with preferences from click%2Dthrough data.

Applying ranking SVM in query relaxation.

Oracle at TREC 10 Filtering and Question%2DAnswering.
Oracle’s objective in TREC-10 was to study the behavior of Oracle information retrieval in previously unexplored application areas. The software used was Oracle9i Text[1], Oracle’s full-text retrieval engine integrated with the Oracle relational database management system, and the Oracle PL/SQL procedural programming language. Runs were submitted in filtering and Q/A tracks. For the filtering track we submitted three runs, in adaptive filtering, batch filtering and routing. By comparing the TREC results, we found that the concepts (themes) extracted by Oracle Text can be used to aggregate document information content to simplify statistical processing. Oracle&#039;s Q/A system integrated information retrieval (IR) and information extraction (IE). The Q/A system relied on a combination of document and sentence ranking in IR, named entity tagging in IE and shallow parsing based classification of questions into pre-defined categories. 1. Filtering based on Theme Signature As a first time filtering track participant, Oracle submitted runs for adaptive filtering, batch filtering and routing this year. Only linear-utility optimized runs were submitted for adaptive filtering and batch filtering. The filtering system is built based on the Oracle 9i database with PL/SQL- an Oracle supported database access language. Since the routing sub-task outputs the top 1000 ranked documents per category, and the training process and similarity score calculation algorithm are the same for batch filtering and routing, we will focus our discussion on batch filtering and adaptive filtering. The filtering system can be divided into three parts based on functionality:
