Microsoft Technology Licensing, LLC

Linkedin Corporation

Shakti Dhirendraji Sinha

Dhruv Arya

Viet Thuc Ha

Satya Kanduri

Ganesh Venkataraman

Abhishek Gupta

Xianren Wu

Asif Mansoor Ali Makhani

Deepak Agarwal

Ye Xu

Eric Huang

Yan Yan

Vijay K Dialani

Abhimanyu Lad

Benjamin Hoan Le

David Golland

Alexandre Patry

Patrick A Chase

Jia Li

Lin Guo

Ramesh Dommeti

Senthil Sundaram

Ashley Woodman Hall

Mario Sergio Rodriguez

Xiaochuan Ni

Shan Zhou

Bradley Scott Mauney

Jiong Wang

Rahim Daya

Zian Yu

Jordan Anthony Saints

Timothy Patrick Jordt

Gregory Alan Walloch

Zachary Tyler Piepmeyer

Xufei Wang

Scott Blackburn

Viet-Ha Thuc

A system and method for calculating quality score for digital content are provided. In example embodiments, a first graph is generated comprising a user node and an article node, the user node corresponds to a user and the article node corresponds to an article. An edge is generated between the user node and the article node in the first graph based on a first action. A second graph is generated comprising the user node and the article node. An edge is generated between the user node and the article node in the second graph based on a second action type. A first authority score is calculated for the article node within the first graph. A second authority score is calculated for the article node within the second graph. A quality score is calculated for the article.

Automated measurement of content quality

In an example embodiment, one or more specified ideal candidates are used to perform a search in a database. One or more attributes are extracted from one or more ideal candidate member profiles. A search query is then generated based on the extracted one or more attributes. Then, a search is performed on member profiles in the social networking service using the generated search query, returning one or more result member profiles.

Query building for search by ideal candidates

In an example embodiment, one or more query terms are obtained. For each of the one or more query terms, a standardized entity taxonomy is searched to locate a standardized entity that most closely matches the query term. A confidence score is calculated for the query term-standardized entity pair for the standardized entity that most closely matches the query term. In response to a determination that the confidence score transgresses a threshold, the query term is associated with an entity identification corresponding to the standardized entity that most closely matches the query term. One or more query rewriting rules corresponding to an entity type of the standardized entity having the entity identification are obtained. The one or more query rewriting rules are executed to rewrite the first query such that the rewritten query, when performed on a data source, returns fewer search results than the first query would have.

Entity based query filtering

In an example embodiment, one or more query terms are obtained. Then, for each of the one or more query terms, a standardized entity taxonomy is searched to locate a standardized entity that most closely matches the query term, with the standardized entity taxonomy comprising an entity identification for each of a plurality of different standardized entities. A confidence score is then calculated for the query term-standardized entity pair for the standardized entity that most closely matches the query term, and the query term is tagged with the entity identification corresponding to the standardized entity that most closely matches the query term and the calculated confidence score.

Entity based search retrieval and ranking

In an example embodiment, usage information is used to calculate one or more individual document historical information-deficient metrics (IDHIDMs) by combining values for the one or more metrics from multiple documents within the document corpus that share an identical combination of entities of the one or more entity types. A search query is segmented into a plurality of segments, wherein at least one of the plurality of segments is tagged as a first entity type and at least one of the plurality of segments is tagged as a second entity type. At least one for a combination of entities matching the tagged segments is used to rank one or more retrieved documents responsive to the query.

Entity-faceted historical click-through-rate

A system and method for generating modifiers for updated search queries are provided. In example embodiments, metadata is accessed, the metadata corresponds to search results of an input query and comprising a plurality of candidate modifiers. A score is calculated for each candidate based on a relevance value that indicates the correlation between a candidate modifier and the input query. A list of top number of candidate modifiers is generated based on the score of the candidate modifier transgressing a first threshold. A uniqueness score is calculated for combination pairs of candidate modifiers within the list, the uniqueness score being used to eliminate candidate modifiers. The list of top number of candidate modifiers is presented, at a user interface, according to a ranked order based on the score.

Generating modifiers for updating search queries

A system and method for automatic topic tagging are provided. In example embodiments, input content is received, the content includes a plurality of terms. Term vectors are generated from the plurality of terms. Candidate topics are identified to assigned to the plurality of terms. Topics are assigned to the received content from the identified candidate topics.

Automatic content tagging

In an example embodiment, one or more sample ideal candidate member profiles in a social networking service are obtained, as well as one or more sample search result member profiles in the social networking service. Then, for each unique pair of sample ideal candidate member profile and sample search result member profile, a label is generated using a score generated from log information of the social networking service, the log information including records of communications between a searcher and members of the social networking service, the score being higher if the searcher communicated with both the member corresponding sample ideal candidate member profile and the member corresponding to the sample search result member profile in a same search session. The generated labels are fed into a machine learning algorithm to train a combined ranking model used to output ranking scores for search result member profiles.

Generation of training data for ideal candidate search ranking model

A user submits a job search query in an online social networking system. The online social networking system calculates a score based on the similarity between the job search query and the profile of the user. When the score transgresses a threshold, the job search query is enhanced by adding data from the profile of the user to the job search query. The job search query is then used to search for, identify, and display jobs in the online social networking system.

Personalized job search and recommendations using job seeker behavioral features

System and techniques for facet expansion are described herein. A user interface element may be presented on facet selection portion of a search result display including search results. Here, the user interface element is arranged to accept user input of a facet. Partial user input for a facet may be received. A peer entity to an entity corresponding to the facet may be obtained. A peer facet may be presented in a suggestion element in the facet selection portion in response to receiving the partial user input.

System for facet expansion

In an example, a plurality of member profiles in a social networking service are obtained, each member profile identifying a member and listing one or more skills the corresponding member has explicitly added to the member profile, the one or more skills indicating a proficiency by the member in the corresponding skill. A members-skills matrix is formed, wherein each cell in the matrix is assigned a value based on whether the corresponding member has the corresponding skill. The dot product of the members matrix and the skills matrix is then computed and used to identify one or more latent skills of a first member of the social networking service. Then a first digitally stored member profile is augmented with the one or more latent skills by combining the one or more latent skills with explicit skills for purposes of one or more searches that utilize member skills as an input variable.

Entity-aware features for personalized job search ranking

In an example embodiment, a query for search results is received, the query including at least one value for one facet, a facet defining a categorical dimension for the search results. It is then determined that the facet in the query is exclusive. In response to the determination that the facet is exclusive: for each potential facet different from the facet in the query: for each potential value in the potential facet: conditional entropy gain of the value in the query and the potential value is determined. The potential value in the potential facet that has the highest conditional entropy gain is determined, as is the potential facet with the minimum maximum conditional entropy gain. Then the potential facet with the minimum maximum is input into a machine learning model, causing the machine learning model to output one or more suggested facets to add to the query.

Smart suggestions for query refinements

Techniques for improving search relevance using past searchers' reputation are described. According to various embodiments, a specification of a search query term corresponding to a skill is received from a searcher, in connection with a search for members of an online social networking service having the skill. Thereafter, a list of search results is generated and displayed based on the search query term, each of the search results corresponding to a member of the online social networking service. A subsequent interaction between the searcher and a specific member corresponding to one of the search results is detected. A skill reputation score associated with the searcher and the skill is then accessed. Thereafter, a search relevance score associated with the specific member and the skill is modified based on the skill reputation score associated with the searcher and the skill.

Search relevance using past searchers' reputation

Growing community of inventors

Sunnyvale, CA, United States of America

Shakti Dhirendraji Sinha