Microsoft Technology Licensing, LLC

Kinjal Basu

Shaunak Chatterjee

Ankan Saha

Mark Hull

Shipeng Yu

Jinyun Yan

Birjodh S Tiwana

Guanchao Wang

Yiping Yuan

Onkar Anant Dalal

Pratik Daga

Siva Visakan Sooriyan

Sowmitra Thallapragada

Jesse Jyh-Cherng Hsia

Sammy Shreibati

Matthew Hsing Hung Walker

Revant Kumar

Haiyang Liu

Vivian Mak Urata

Josh Fleming

Chengming Jiang

Zhiyuan Xu

Manoj Sivakumar

Yunbo Ouyang

Michael Joshua Aft

Varun Bharill

Techniques are provided for automatically tuning a parameter in a layered model framework. One or more machine learning techniques are used to train multiple versions of a first model that includes a first version and a second version. A second model is stored that includes a parameter and accepts, as input, output from the first model. Multiple parameter values of the parameter are tested when processing content requests using the first and second versions of the first model. A strict subset of the plurality of parameter values are selected for the parameter of the second model, such that processing a first subset of the content requests using the first version of the first model results in a first value of a particular metric that matches a second value of the particular metric resulting from processing a second subset of the content requests using the second version of the first model.

Automatically tuning parameters in a layered model framework

Described herein is a contextual contact recommendation or suggestion service and system. The service, which, in some embodiments, is integrated with a social networking service and/or an instant messaging service, takes as input a first parameter that identifies a member of the social networking service, and a second parameter that defines a context (e.g., a web page that is being viewed by the member. The service, based in part on the context, computes a ranked list of members to populate a contextual contact list, thereby recommending or suggesting contacts, with whom the member might be interested in initiating, or continuing, a conversation, based on the context of the member's current web browsing session. Optionally, the service may take as input a third parameter, defining a use case, such that the recommendation algorithm can be customized by use case.

Techniques for presenting a contextual contact list

Systems and methods determine optimized hyperparameter values for one or more machine-learning models. A sample training data set from a larger corpus of training data is obtained. Initial hyperparameter values are then randomly selected. Using the sample training data set and the randomly chosen hyperparameter values, an initial set of performance metric values are obtained. Maximized hyperparameter values are then determined from the initial set of hyperparameter values based on the corresponding performance metric value. A larger corpus of training data is then evaluated using the maximized hyperparameter values and the corresponding machine-learning model, which yields another corresponding set of performance metric values. The maximized hyperparameter values and their corresponding set of performance metric values are then merged with the prior set of hyperparameter values. The foregoing operations are performed iteratively until it is determined that the hyperparameter values are converging to a particular value.

Large-scale automated hyperparameter tuning

A system, a machine-readable storage medium storing instructions, and a computer-implemented method are described herein are directed to a Content Optimization Engine that determines a display probability for each content item in a set of content items. Each respective display probability corresponds to a given content item's probability of display in a specific content slot of a plurality of content slots in a social network feed of a target member account in a social network service. The Content Optimization Engine calculates a selection probability for each content item in an ordered set of the content items, based on each display probability and a set of interaction effects. The Content Optimization Engine causes display of the ordered set of content items in the target member account's social network feed based on satisfaction of the first and second targets.

Constrained multi-slot optimization for ranking recommendations

A unified notification platform for offline creation and distribution of notification content from a variety of data sources is described. The notification platform provides data adaptors that are reusable for generating notifications of different types, specifically, for generating notifications of different types that have features that have meaning across different notification types such that these features can be used to generate comparable relevance scores with respect to candidate profiles. The relevance score calculated for a notification with respect to a member profile is used to determine whether the notification is to be presented to the member represented by the member profile.

Content generation and targeting

A system, a machine-readable storage medium storing instructions, and a computer-implemented method are described herein are directed to Large Scale Optimizing Engine. The Large Scale Optimizing Engine determines a probability, for each content item in a set of content items, of the respective member account performing a content item action. The Large Scale Optimizing Engine identifies a select content item from the set of content items based on determining display of the select content item will meet a first and second target. The Large Scale Optimizing Engine causes display of the select content item in a content slot in the respective member account's social network feed based on satisfaction of the first and second targets.

Large scale multi-objective optimization

Growing community of inventors

Stanford, CA, United States of America

Kinjal Basu