-
公开(公告)号：US20210326751A1
公开(公告)日：2021-10-21
申请号：US16882296
申请日：2020-05-22
发明人： Xiaodong Liu , Hao Cheng , Yu Wang , Jianfeng Gao , Weizhu Chen , Pengcheng He , Hoifung Poon
IPC分类号： G06N20/00 , G06N3/08 , G06F40/284 , G06K9/62
摘要： This document relates to training of machine learning models. One example method involves providing a machine learning model having one or more mapping layers. The one or more mapping layers can include at least a first mapping layer configured to map components of pretraining examples into first representations in a space. The example method also includes performing a pretraining stage on the one or more mapping layers using the pretraining examples. The pretraining stage can include adding noise to the first representations of the components of the pretraining examples to obtain noise-adjusted first representations. The pretraining stage can also include performing a self-supervised learning process to pretrain the one or more mapping layers using at least the first representations of the training data items and the noise-adjusted first representations of the training data items.
-
公开(公告)号：US10579423B2
公开(公告)日：2020-03-03
申请号：US15943206
申请日：2018-04-02
发明人： Jinchao Li , Yu Wang , Karan Srivastava , Jianfeng Gao , Prabhdeep Singh , Haiyuan Cao , Xinying Song , Hui Su , Jaideep Sarkar
摘要： Generally discussed herein are devices, systems, and methods for scheduling tasks to be completed by resources. A method can include identifying features of the task, the features including a time-dependent feature and a time-independent feature, the time-dependent feature indicating a time the task is more likely to be successfully completed by the resource, converting the features to feature values based on a predefined mapping of features to feature values in a first memory device, determining, by a gradient boost tree model and based on a first current time and the feature values, a likelihood the resource will successfully complete the task, and scheduling the task to be performed by the resource based on the determined likelihood.
-
公开(公告)号：US20190303773A1
公开(公告)日：2019-10-03
申请号：US15941236
申请日：2018-03-30
发明人： Chi-Yi Kuan , Shen Huang , Yu Wang , Yongzheng Zhang , Paul Ko , Shady Elasra , Fanbin Bu
摘要： Techniques are provided for predicting talent flow to and/or from a geographical region. In one technique, multiple entity profiles are stored and analyzed to generate training data that is labeled indicating whether a corresponding entity has moved to or moved from a region. A machine-learned prediction model is generated or trained based on the training data. Using the machine-learned prediction model, a prediction is made whether, for each entity corresponding to another entity profile, that entity will move to or move from a particular geographic region. Based on multiple predictions, a number of entities that are predicted to move to or move from the particular geographic region is determined. Talent flow data that is based on the number of entities is presented on a computer display.
-
公开(公告)号：US11966769B2
公开(公告)日：2024-04-23
申请号：US17601635
申请日：2019-05-23
发明人： Eric Ray Hotinger , Bin Du , Sajay Antony , Steven M. Lasker , Siva Garudayagari , Dongjiang You , Yu Wang , Samarth Shah , Brian Timothy Goff , Shiwei Zhang
CPC分类号： G06F9/45558 , G06F8/63 , G06F2009/45562 , G06F2009/45579
摘要： Computing system enhancements make container instantiation faster, reduce layer content storage demands, and make more container image formats available. A container instantiation location sends a container image pull request to a container registry, receives an image manifest, sends a layer mount request to the registry instead of a layer content download request, receives a layer mount, optionally repeats for additional layers, creates a union file system spanning the layers, and launches a container process based on the union file system without first downloading all the layer content. Inefficiencies and technical limitations of some other approaches are avoided, such as loopback mounts for snapshot expansion, creation or transmission of extra snapshots or extra container image clones, cluttering layer content with virtual machine settings, container system vendor lock-in, lack of container instantiation at a local system due to insufficient local storage, and lack of syscall optimization due to storage driver plugin usage.
-
公开(公告)号：US11238352B2
公开(公告)日：2022-02-01
申请号：US15941236
申请日：2018-03-30
发明人： Chi-Yi Kuan , Shen Huang , Yu Wang , Yongzheng Zhang , Paul Ko , Shady Elasra , Fanbin Bu
摘要： Techniques are provided for predicting talent flow to and/or from a geographical region. In one technique, multiple entity profiles are stored and analyzed to generate training data that is labeled indicating whether a corresponding entity has moved to or moved from a region. A machine-learned prediction model is generated or trained based on the training data. Using the machine-learned prediction model, a prediction is made whether, for each entity corresponding to another entity profile, that entity will move to or move from a particular geographic region. Based on multiple predictions, a number of entities that are predicted to move to or move from the particular geographic region is determined. Talent flow data that is based on the number of entities is presented on a computer display.
-
公开(公告)号：US11068304B2
公开(公告)日：2021-07-20
申请号：US16285170
申请日：2019-02-25
发明人： Jinchao Li , Xinying Song , Ah Young Kim , Haiyuan Cao , Yu Wang , Hui Su , Shahina Ferdous , Jianfeng Gao , Karan Srivastava , Jaideep Sarkar
IPC分类号： G06F9/48 , G06F9/46 , H04M3/52 , H04M3/51 , G06N20/00 , G06Q10/04 , G06Q10/06 , G06Q10/10 , G06F9/50 , G06K9/62 , H04M3/523 , G06N7/00 , G06N3/08 , G06N20/20
摘要： Systems and methods are disclosed for intelligent scheduling of calls to sales leads, leveraging machine learning (ML) to optimize expected results. One exemplary method includes determining, using a connectivity prediction model, call connectivity rate predictions; determining timeslot resources; allocating, based at least on the call connectivity rate predictions and timeslot resources, leads to timeslots in a first time period; determining, within a timeslot and using a lead scoring model, lead prioritization among leads within the timeslot; configuring, based at least on the lead prioritization, the telephone unit with lead information for placing a phone call; and applying a contextual bandit (ML) process to update the connectivity prediction model, the lead scoring model, or both. During subsequent time periods, the updated connectivity prediction and lead scoring models are used, thereby improving expected results over time.
-
公开(公告)号：US20200349605A1
公开(公告)日：2020-11-05
申请号：US16401832
申请日：2019-05-02
发明人： Sahin C. Geyik , Florian Raudies , Xi Chen , Yu Wang , Wen Pu
摘要： The disclosed embodiments provide a system for performing calibration of response rates. During operation, the system obtains a position of a content item in a ranking of content items generated for delivery to a member of an online system and a predicted response rate by the member to the content item. Next, the system determines an updated response rate by the member to the content item based on the position of the content item in the ranking and dimensions associated with the predicted response rate and the ranking. The system then outputs the updated response rate for use in managing delivery of the content item.
-
公开(公告)号：US20190303197A1
公开(公告)日：2019-10-03
申请号：US15943206
申请日：2018-04-02
发明人： Jinchao Li , Yu Wang , Karan Srivastava , Jianfeng Gao , Prabhdeep Singh , Haiyuan Cao , Xinying Song , Hui Su , Jaideep Sarkar
摘要： Generally discussed herein are devices, systems, and methods for scheduling tasks to be completed by resources. A method can include identifying features of the task, the features including a time-dependent feature and a time-independent feature, the time-dependent feature indicating a time the task is more likely to be successfully completed by the resource, converting the features to feature values based on a predefined mapping of features to feature values in a first memory device, determining, by a gradient boost tree model and based on a first current time and the feature values, a likelihood the resource will successfully complete the task, and scheduling the task to be performed by the resource based on the determined likelihood.
-
公开(公告)号：US10366131B2
公开(公告)日：2019-07-30
申请号：US15161050
申请日：2016-05-20
发明人： Ryen W. White , Zijian Zheng , An Yan , Xiao Huang , Yu Wang
IPC分类号： G06F16/2457 , G06F16/9535 , G06F16/9537 , G06F9/48 , G06N5/02 , G06N7/00 , G06F17/30
摘要： 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.
-
公开(公告)号：US20240013055A1
公开(公告)日：2024-01-11
申请号：US18373051
申请日：2023-09-26
发明人： Xiaodong Liu , Hao Cheng , Yu Wang , Jianfeng Gao , Weizhu Chen , Pengcheng He , Hoifung Poon
CPC分类号： G06N3/084 , G06N20/00 , G06N3/08 , G06N3/088 , G06V10/82 , G06F18/24 , G06V10/7784 , G06F40/284
摘要： This document relates to training of machine learning models. One example method involves providing a machine learning model having one or more mapping layers. The one or more mapping layers can include at least a first mapping layer configured to map components of pretraining examples into first representations in a space. The example method also includes performing a pretraining stage on the one or more mapping layers using the pretraining examples. The pretraining stage can include adding noise to the first representations of the components of the pretraining examples to obtain noise-adjusted first representations. The pretraining stage can also include performing a self-supervised learning process to pretrain the one or more mapping layers using at least the first representations of the training data items and the noise-adjusted first representations of the training data items.
-
-
-
-
-
-
-
-
-
搜索结果
20 条记录 (耗时 0.031 秒)
