公开(公告)号：US11461421B2

公开(公告)日：2022-10-04

申请号：US17136864

申请日：2020-12-29

申请人： Microsoft Technology Licensing, LLC

发明人： Yiming Wang ,  Xiao Yan ,  Lin Zhu ,  Jaewon Yang ,  Yanen Li ,  Jacob Bollinger

IPC分类号： G06F16/9535 ,  G06N3/04 ,  G06N20/20 ,  G06F16/36

摘要： Techniques for ranking skills using an ensemble machine learning approach are described. The outputs of two heterogenous, machine-learned models are combined to rank a set of skills that may be possessed by an end-user of an online service. Some subset of the highest-ranking skills is then presented to the end-user with a recommendation that the skills be added to the end-user's profile. The ensemble learning technique involves a concept referred to as “boosting”, in which a weaker performing model is enhanced (e.g., “boosted”) by a stronger performing model, when ranking the set of skills. Accordingly, by using a combination of models, better results are achieved than might be with either one of the individual models alone. Furthermore, the approach is scalable in ways that cannot be achieved with heuristic-based approaches.

公开(公告)号：US20220207099A1

公开(公告)日：2022-06-30

申请号：US17136864

申请日：2020-12-29

申请人： Microsoft Technology Licensing, LLC

发明人： Yiming Wang ,  Xiao Yan ,  Lin Zhu ,  Jaewon Yang ,  Yanen Li ,  Jacob Bollinger

IPC分类号： G06F16/9535 ,  G06N3/04 ,  G06F16/9536 ,  G06F16/36 ,  G06N20/20

摘要： Techniques for ranking skills using an ensemble machine learning approach are described. The outputs of two heterogenous, machine-learned models are combined to rank a set of skills that may be possessed by an end-user of an online service. Some subset of the highest-ranking skills is then presented to the end-user with a recommendation that the skills be added to the end-user's profile. The ensemble learning technique involves a concept referred to as “boosting”, in which a weaker performing model is enhanced (e.g., “boosted”) by a stronger performing model, when ranking the set of skills. Accordingly, by using a combination of models, better results are achieved than might be with either one of the individual models alone. Furthermore, the approach is scalable in ways that cannot be achieved with heuristic-based approaches.

公开(公告)号：US10412189B2

公开(公告)日：2019-09-10

申请号：US14953943

申请日：2015-11-30

申请人： Microsoft Technology Licensing, LLC

发明人： Jacob Bollinger ,  David Hardtke ,  Bo Zhao

IPC分类号： H04L29/08 ,  G06F16/901 ,  H04W4/21

摘要： This disclosure is directed to determining various economic graph indices and, in particular, to systems and methods that leverage a graph analytic engine and framework to determine values assigned to graph nodes extracted from one or more member profiles, and visualizing said values to correlate skills, geographies, and industries. The disclosed embodiments include a client-server architecture where a social networking server has access to a social graph of its social networking members. The social networking server includes various modules and engines that import the member profiles and then extracts certain defined attributes from the member profiles, such as employer (e.g., current employer and/or past employers), identified skills, educational institutions attended, and other such defined attributes. Using these attributes as nodes, the social networking server constructs a graph using various graph processing techniques. The resulting graph is then used to correlate and rank the various attributes that define the graph.

公开(公告)号：US11604990B2

公开(公告)日：2023-03-14

申请号：US16902587

申请日：2020-06-16

申请人： Microsoft Technology Licensing, LLC

发明人： Xiao Yan ,  Wenjia Ma ,  Jaewon Yang ,  Jacob Bollinger ,  Qi He ,  Lin Zhu ,  How Jing

IPC分类号： G06N3/08 ,  G06F9/48 ,  H04L67/50 ,  G06F30/27 ,  G06N3/02

摘要： In an example embodiment, a framework to infer a user's value for a particular attribute based upon a multi-task machine learning process with uncertainty weighting that incorporates signals from multiple contexts is provided. In an example embodiment, the framework aims to measure a level of a user attribute under a certain context. Rather than attempting to devise a universal, one-size-fits-all value for the attribute, the framework acknowledges that the user's value for that attribute can vary depending on context and factors in the context under which the user's attribute levels are measured. Multiple contexts are defined depending on different situations where users and entities such as companies and organizations need to evaluate user attribute levels. Signals for attribute levels are then collected for each context. Machine learning models are utilized to estimate attribute values for different contexts. Multi-task deep learning is used to level attributes from different contexts.