Abstract: The subject disclosure is directed towards building one or more context and query models representative of users' search interests based on their logged interaction behaviors (context) preceding search queries. The models are combined into an intent model by learning an optimal combination (e.g., relative weight) for combining the context model with a query model for a query. The resultant intent model may be used to perform a query-related task, such as to rank or re-rank online search results, predict future interests, select advertisements, and so forth.

Abstract: A “Cross-Lingual Unified Relevance Model” provides a feedback model that improves a machine-learned ranker for a language with few training resources, using feedback from a more complete ranker for a language that has more training resources. The model focuses on linguistically non-local queries, such as “world cup” (English language/U.S. market) and “copa mundial” (Spanish language/Mexican market), that have similar user intent in different languages and markets or regions, thus allowing the low-resource ranker to receive direct relevance feedback from the high-resource ranker. Among other things, the Cross-Lingual Unified Relevance Model differs from conventional relevancy-based techniques by incorporating both query- and document-level features.

Abstract: Measuring duplication in search results is described. In one example, duplication between a pair of results provided by an information retrieval system in response to a query is measured. History data for the information retrieval system is accessed and query data retrieved, which describes the number of times that users have previously selected either or both of the pair of results, and a relative presentation sequence of the pair of results when displayed at each selection. From the query data, a fraction of user selections is determined in which a predefined combination of one or both of the pair of results were selected for a predefined presentation sequence. From the fraction, a measure of duplication between the pair of results is found. In further examples, the information retrieval system uses the measure of duplication to determine an overall redundancy value for a result set, and controls the result display accordingly.

Abstract: A system that facilitates ranking search results returned by a search engine in response to receipt of a query is described herein. The system includes a receiver component that receives categorical metadata pertaining to an item and categorical metadata pertaining to the query and a computation component that computes at least one of a document feature pertaining to the item, a query feature pertaining to the query, or a document-query feature pertaining to the item and the query based at least in part upon one or more of the categorical metadata pertaining to the item or the categorical metadata pertaining to the query. The system also includes a ranker component that selectively places the item in a particular location in a sequence of items based at least in part upon the at least one of the document feature, the query feature, or the document-query feature.

Abstract: The hierarchical approach may start at the bottom of the hierarchy. As it moves up the hierarchy, knowledge from children and cousins is used to classify items at the parent. In addition, knowledge of improper classifications at a low level are raised to a higher level to create new rules to better identify mistaken classifications at a higher level. Once the top of the hierarchy is reached, a top down approach is used to further refine the classification of items.

Abstract: The invention applies a probabilistic approach to combining evidence regarding the correct classification of items. Training data and machine learning techniques are used to construct probabilistic dependency models that effectively utilize evidence. The evidence includes the outputs of one or more classifiers and optionally one or more reliability indicators. The reliability indicators are, in a broad sense, attributes of the items being classified. These attributes can include characteristics of an item, source of an item, and meta-level outputs of classifiers applied to the item. The resulting models include meta-classifiers, which combine evidence from two or more classifiers, and tuned classifiers, which use reliability indicators to inform the interpretation of classical classifier outputs. The invention also provides systems and methods for identifying new reliability indicators.