Microsoft Technology Licensing, LLC

Mikhail Bilenko

Ran Gilad-Bachrach

Hoyt Adam Koepke

Miles Arthur Munson

Christopher A Meek

Ryen William White

Hugh Evan Williams

Ewa Dominowska

Matthew R Richardson

Nicholas Eric Craswell

Michael Maxwell Cameron

Wei Yu

Alice Zheng

Amir Jonatan Padovitz

Pengtao Xie

Mikhail Aleksandrovich Royzner

Thomas William Finley

Hrishikesh Bal

Scott Schult

Dima Suponau

Deepak Bapna

Anthony Crispo

Arunesh Gupta

Marty Kauhanen

A training system is described herein for generating a prediction model that relies on a feature space with reduced dimensionality. The training system performs this task by producing partitions, each of which corresponds to a subset of aspect values (where each aspect value, in turn, may correspond to one or more attribute values). The training system then produces instances of statistical information associated with the partitions. Each instance of statistical information therefore corresponds to feature information that applies to a plurality of aspect values, rather than a single aspect value. The training system then trains the prediction model based on the feature information. Also described herein is a prediction module that uses the prediction model to make predictions in various online contexts.

Predicting behavior using features derived from statistical information

In one embodiment, an evolving advertising system automatically optimizes internet advertising. A data storage unitmay store an evolving advertisement unitwith an advertisement characteristic according to an initial configuration parameter. A communication interfacemay transmit the evolving advertisement unitas part of a primary website. A processormay alter the evolving advertisement unitautomatically upon a trigger event by changing the advertisement characteristic to follow an automatically generated configuration parameter to optimize an advertisement performance metric.

Learning display parameters to maximize advertising revenue

Various technologies described herein pertain to estimating predictive accuracy gain of a potential feature added to a set of features, wherein an existing predictor is trained on the set of features. Outputs of the existing predictor for instances in a dataset can be retrieved from a data store. Moreover, a predictive accuracy gain estimate of a potential feature added to the set of features can be measured as a function of the outputs of the existing predictor for the instances in the dataset. The predictive accuracy gain estimate can be measured without training an updated predictor on the set of features augmented by the potential feature.

Estimation of predictive accuracy gains from added features

Embodiments described herein are directed to methods and systems for performing neural network computations on encrypted data. Encrypted data is received from a user. The encrypted data is encrypted with an encryption scheme that allows for computations on the ciphertext to generate encrypted results data. Neural network computations are performed on the encrypted data, using approximations of neural network functions to generate encrypted neural network results data from encrypted data. The approximations of neural network functions can approximate activation functions, where the activation functions are approximated using polynomial expressions. The encrypted neural network results data are communicated to the user associated with the encrypted data such that the user decrypts the encrypted data based on the encryption scheme. The functionality of the neural network system can be provided using a cloud computing platform that supports restricted access to particular neural networks.

Neural networks for encrypted data

Technologies pertaining to tuning a hyper-parameter configuration of a learning algorithm are described. The learning algorithm learns parameters of a predictive model based upon the hyper-parameter configuration. Candidate hyper-parameter configurations are identified, and statistical hypothesis tests are undertaken over respective pairs of candidate hyper-parameter configurations to identify, for each pair of candidate hyper-parameter configurations, which of the two configurations is associated with better predictive performance. The technologies described herein take into consideration the stochastic nature of training data, validation data, and evaluation functions.

Tuning hyper-parameters of a computer-executable learning algorithm

Technologies pertaining to computing a tiering policy that defines how digital items are desirable stored across a plurality of different storage tiers are described herein. A data repository that comprises data that is indicative of historic user interaction with a search engine is accessed. Subsequently, a tiering policy for digital items that are retrievable by way of the search engine is computed based at least in part upon the data that is indicative of the historic user interaction with the search engine. Retrieval times for digital items in the data storage tiers differ across the data storage tiers.

Determining quality of tier assignments

Described herein is a method that includes receiving user history data and generating an indication of quality of a tier assignment used to store searchable digital items in a tiered storage system, wherein the indication is based at least in part upon a subset of the user history data. Also described herein is a system that includes a receiver component that receives user history data. The system further includes a quality indicator component that determines an indication of quality of a tier assignment used to store digital items that are retrievable by way of querying, wherein the quality indicator component generates the indication based at least in part upon a subset of the user history data and the tier assignment indicates where digital items are to be stored in a tiered storage system.

A two-phase process quickly and accurately identifies representations of the same items within a collection of item representations. In the first phase, referred to as a 'blocking phase,' frequency information indicating the frequency with which terms appear within the collection of item representations is used to quickly identify 'candidate pairs' (i.e., pairs of item representations that have a relatively high probability of matching). The blocking phase results in a reduced subset of the data for further analysis during the second phase. In the second phase, referred to as a &#x201c;matching phase,&#x201d; the candidate pairs are analyzed using fuzzy matching functions to accurately identify &#x201c;matching pairs&#x201d; (i.e., representations of the same items).

Large scale item representation matching

It has been found that user navigation that follows search engine interactions provides implicit endorsement of resources (such as web resources) that are preferred by users, and which may be particularly valuable for exploratory search tasks. Thus, a combination of past searching and browsing user behavior is analyzed to identify additional information that augments search results delivered by a search engine. The additional information may include a display of hyperlinks to locations which are derived from the past searching and browsing user behavior, given a specific input query. The additional information may be provided to supplement web search results.

Growing community of inventors

Bellevue, WA, United States of America

Mikhail Bilenko