Microsoft Technology Licensing, LLC

Zijian Zheng

Pyungchul Kim

Ryen William White

Yu Wang

Peng Yu

An Yan

Nancy Ellen Jacobs

Frederic H Behr, Jr

Steven Fox

Shengquan Yan

Xiao Huang

Li Deng

Jianfeng Gao

Kuansan Wang

Xiaodong He

Wei Guang Lu

Sanjeev Katariya

LuMing Wang

Jianshu Chen

Zhenbin Xu

Prabhdeep Singh

Randall Kern

Sin Lew

Mingyang Zhao

Mark B Mydland

Yu Fan Chen

Robert C Wang

Toby H Walker

Miriam Rosenberg

Xian Zhang

Guixi Zou

Yin He

Michael D Hintze

Xavier Drudis Rius

Kimberly J Howell

Simone Maillard

Brendan W L Clement

Bochen Jin

Frederic Behr

Sandeep P Solanki

Hanjun Xian

Yilei Zhang

Michael Kinoti

Architecture that decomposes of one or more monolithic data concepts into atomic concepts and related atomic concept dependencies, and provides streaming data processing that processes individual or separate (atomic) data concepts and defined atomic dependencies. The architecture can comprise data-driven data processing that enables the plug-in of new data concepts with minimal effort. Efficient processing of the data concepts is enabled by streaming only required data concepts and corresponding dependencies and enablement of the seamless configuration of data processing between stream processing systems and batch processing systems as a result of data concept decomposition. Incremental and non-incremental metric processing enables realtime access and monitoring of operational parameters and queries.

Stream processing in search data pipelines

A processing unit can acquire datasets from respective data sources, each having a respective unique data domain. The processing unit can determine values of a plurality of features based on the plurality of datasets. The processing unit can modify input-specific parameters or history parameters of a computational model based on the values of the features. In some examples, the processing unit can determine an estimated value of a target feature based at least in part on the modified computational model and values of one or more reference features. In some examples, the computational model can include neural networks for several input sets. An output layer of at least one of the neural networks can be connected to the respective hidden layer(s) of one or more other(s) of the neural networks. In some examples, the neural networks can be operated to provide transformed feature value(s) for respective times.

Training and operation of computational models

The concepts relate to task completion and specifically to aiding a user to complete an unfinished task at a subsequent time and/or on another device. One example can identify that a user is working on a task on a computing device associated with the user. In an instance when the user stops using the computing device without completing the task, the example can predict a likelihood that the user will subsequently resume the task on a second computing device associated with the user. In an instance where the likelihood exceeds a threshold, the example can attempt to aid the user in completing the task on the second computing device.

Task completion

The ranking quality of a ranked list may be evaluated. In an example embodiment, a method is implemented by a system to access log data, ascertain which entries of a ranked list are skipped, and determine a ranking quality metric from the skipped entries. More specifically, log data that reflects user interactions with a ranked list having multiple entries is accessed. The user interactions include at least indications of which of the multiple entries are selected entries. It is ascertained which entries of the multiple entries of the ranked list are skipped entries based on the selected entries. The ranking quality metric for the ranked list is determined responsive to the skipped entries.

Evaluating the ranking quality of a ranked list

Systems and methods are provided for using monads to facilitate complex computation tasks in a cloud computing environment. In particular, monads can be employed to facilitate creation and execution of data mining jobs for large data sets. Monads can allow for improved error handling for complex computation tasks. Monads can also facilitate identification of opportunities for improving the efficiency of complex computations.

Monad based cloud computing

Methods for updating an information retrieval system are disclosed. In one embodiment, search terms affiliated with mappings or associations that represent a connection of relevancy between a query and an asset are pushed as content updates to a client system (e.g., as new updates or utilized to replace older data). The search terms are inserted (e.g., inserted as metadata) into corresponding content (the content associated with the asset). In this manner, content-searching data can be updated (e.g., remotely updated) as frequently as desired, even periodically, or selectively as new manually and/or automatically derived data becomes available. In another embodiment, the update data is already built into the content when it is delivered to a client machine. Other disclosed embodiments pertain to methods for generating a data mining classification model that is a blended representation of associations (e.g., query-asset associations) having different characteristics and/or different originating sources.

Generation of a blended classification model

Assigning session identifications to log entries and generating anonymous log entries are provided. In order to balance users' privacy concerns with the need for analysis of the log entries to provide high quality search results, non-user-specific data fields, such as a user's location (e.g., city, state, and latitude/longitude) and connection speed, are inserted into the log entries, and user-specific data fields, such as the IP address and cookie identifications, are deleted from the log entries. In addition or alternatively, prior to anonymization of the log entries, session identifications are assigned to identified groups of log entries. The groups are identified based on factors such as the user's identification, the IP address, the time of search, and differences between the search terms used in the search queries.

Generating anonymous log entries

A system for dynamically updating user accessible features of a software application on a client computer has a user interface, a local usage data file, and a data mining engine. The user interface is adapted to receive operator inputs. The local usage data file is adapted to store usage information corresponding to the operator inputs. The data mining engine is adapted to process the stored usage information and to generate local adjustments to a user interface of the software application based on the operator inputs. In one embodiment, a server is adapted to receive usage data from a plurality of application instances on a plurality of client computers and to generate global adjustments based on the received usage data. In one embodiment, the system has a merge feature adapted to blend and resolve conflicts between local and global adjustments to generate an interface adjustment for the user interface.

Adaptive systems and methods for making software easy to use via software usage mining

A search authoring tool for mapping queries to assets for use in an information retrieval system has a query-to-asset mappings database, a search authoring engine, and a search evaluation engine. The query-to-asset mappings database is built from search author input and user mappings. The search authoring engine is adapted to receive operator input and to generate one or more new query-to-asset mappings based on the received operator input. The search evaluation engine is adapted to compare the one or more new query-to-asset mappings to the existing mappings to determine if the one or more new query-to-asset mappings will cause performance degradation of the information retrieval system.

Search authoring metrics and debugging

A search mechanism is evaluated or a relevance model checked using session level and result level diagnostics based on user behavior during a search session with respect to queries entered and user responses to result lists. Tracking occurs of when content desired by a user exists, but is not returned in a search result list when a query is made by the user with intent to find the desired content, when content desired by the user does not exist, when content desired by a user exists, but is not recognized by the user in a result list or is too low in a result list. A user's intent and search context is also taken into consideration when performing search mechanism diagnostics.

Systems and methods for discovery of data that needs improving or authored using user search results diagnostics

The subject invention leverages data logging of responses to diagnostic reports to provide data that can be mined for diagnostic report quality information. Instances of the subject invention provide an initial diagnostic report assessment means to facilitate review by an entity. The entity's responses to the sorted diagnostic reports are logged unobtrusively to create diagnostic report quality data. This data is then analyzed by an analysis means that can then adjust the assessment means to improve its performance. In this manner, the performance of the assessment means is increased while reducing the workload of the entity reviewing the diagnostic reports. Other instances of the subject invention facilitate to increase the performance of a diagnostic report generating means as well. Instances of the subject invention can also employ machine learning techniques to facilitate in analyzing the quality data and/or in assessing the diagnostic reports.

Diagnostic report improvement utilizing unobtrusive workflow logging

Growing community of inventors

Bellevue, WA, United States of America

Zijian Zheng