-
1.发明公开
公开(公告)号：EP3866028A2
公开(公告)日：2021-08-18
申请号：EP21166280.4
申请日：2021-03-31
发明人： LIU, Huan , CHENG, Mingquan , CHEN, Kunbin , LIU, Zhun , HE, Bolei , HE, Wei
IPC分类号： G06F16/901 , G06F16/906
摘要： Embodiments of the present disclosure disclose a method and apparatus for constructing a quality evaluation model, an electronic device and a computer-readable storage medium. A specific implementation mode of the method comprises: acquiring (201) samples of knowledge contents; extracting (202) statistical features, semantic features, and image features respectively from the samples of knowledge contents; and constructing (203) a quality evaluation model for knowledge according to the statistical features, the semantic features, and the image features. On the basis of the prior art, this implementation mode additionally uses semantic features and image features of knowledge contents to construct a more accurate quality evaluation model based on multi-dimensional features that characterize the actual quality of a knowledge, which may well discover some brief but very useful summary knowledge in an enterprise and may recommend high-quality knowledge more accurately for employees in the enterprise.
-
公开(公告)号：EP4033734A3
公开(公告)日：2022-12-14
申请号：EP22175399.9
申请日：2022-05-25
发明人： CAO, Shijie , LI, Yanan , HE, Bolei , CHEN, Kunbin , HE, Wei , HE, Feng
IPC分类号： H04L51/04 , H04L51/216 , H04L51/18 , H04L51/52
摘要： A method and an apparatus for generating an account intimacy and a storage medium are provided, which relates to the fields of deep learning, cloud computing and big data. The method includes: obtaining a set of accounts in an instant messaging (IM) group; obtaining a communication frequency between accounts in the set of accounts within a preset time period; generating a communication network graph based on the communication frequency; obtaining an embedding vector of each account output by a graph model, in which the graph model is trained based on the communication network graph; and generating an intimacy between the accounts based on the embedding vectors. In this way, an account intimacy model is generated independently without binding with a downstream business, and the graph model can be effectively used to mine an association between accounts, so as to improve the intimacy generation accuracy.
-
3.发明公开
公开(公告)号：EP3920088A1
公开(公告)日：2021-12-08
申请号：EP21150149.9
申请日：2021-01-05
发明人： LIU, Chengxiang , LIU, Hao , HE, Bolei
摘要： The present disclosure discloses a method and an apparatus for generating a tag of a video. The method includes: obtaining (101, 201, 401, 102, 202, 402) multiple candidate tags and video information of the video; determining (103, 203) first correlation information between the video information and each of the multiple candidate tags; sorting (104, 206, 404) the multiple candidate tags based on the first correlation information to obtain a sort result; and generating (105, 207, 405) the tag of the video based on the sort result. In this way, the integration and utilization for the multiple candidate tags implemented, labeling the video accurately and perfectly, improving the accuracy and the integrity of the tag, and improving the effect of video recommendation.
-
公开(公告)号：EP4033734A2
公开(公告)日：2022-07-27
申请号：EP22175399.9
申请日：2022-05-25
发明人： CAO, Shijie , LI, Yanan , HE, Bolei , CHEN, Kunbin , HE, Wei , HE, Feng
IPC分类号： H04L51/04 , H04L51/216 , H04L51/18 , H04L51/52
摘要： A method and an apparatus for generating an account intimacy and a storage medium are provided, which relates to the fields of deep learning, cloud computing and big data. The method includes: obtaining a set of accounts in an instant messaging (IM) group; obtaining a communication frequency between accounts in the set of accounts within a preset time period; generating a communication network graph based on the communication frequency; obtaining an embedding vector of each account output by a graph model, in which the graph model is trained based on the communication network graph; and generating an intimacy between the accounts based on the embedding vectors. In this way, an account intimacy model is generated independently without binding with a downstream business, and the graph model can be effectively used to mine an association between accounts, so as to improve the intimacy generation accuracy.
-
公开(公告)号：EP3916584A1
公开(公告)日：2021-12-01
申请号：EP21150304.0
申请日：2021-01-05
IPC分类号： G06F16/95
摘要： The present disclosure provides an information processing method, an information processing apparatus, an electronic device and a computer storage medium, relates to a field of information processing. The method includes: obtaining (202) a first content based on a first search keyword indicating a first event and a second search keyword indicating an object related to the first event; obtaining (204) information associated with an attribute of the object from the first content; obtaining (206) a second content based on the first search keyword and a third search keyword indicating a result at least caused by the first event; and generating (208) statistical data associated with the first event based on the information and the second content. Therefore, it is possible to efficiently obtain the content related to the event, and extract the information to generate the statistical data.
-
6.发明公开
公开(公告)号：EP3866028A3
公开(公告)日：2021-10-20
申请号：EP21166280.4
申请日：2021-03-31
发明人： LIU, Huan , CHENG, Mingquan , CHEN, Kunbin , LIU, Zhun , HE, Bolei , HE, Wei
IPC分类号： G06F16/36 , G06F16/901 , G06F16/906
摘要： Embodiments of the present disclosure disclose a method and apparatus for constructing a quality evaluation model, an electronic device and a computer-readable storage medium. A specific implementation mode of the method comprises: acquiring (201) samples of knowledge contents; extracting (202) statistical features, semantic features, and image features respectively from the samples of knowledge contents; and constructing (203) a quality evaluation model for knowledge according to the statistical features, the semantic features, and the image features. On the basis of the prior art, this implementation mode additionally uses semantic features and image features of knowledge contents to construct a more accurate quality evaluation model based on multi-dimensional features that characterize the actual quality of a knowledge, which may well discover some brief but very useful summary knowledge in an enterprise and may recommend high-quality knowledge more accurately for employees in the enterprise.
-
7.发明公开
公开(公告)号：EP4134900A3
公开(公告)日：2023-05-24
申请号：EP22210615.5
申请日：2022-11-30
发明人： XIA, Xiaoling , XU, Wei , CAO, Chengtai , CHEN, Kunbin , HE, Bolei , HE, Wei
IPC分类号： G06Q30/0601 , G06Q30/0251
摘要： The present disclosure provides a method and an apparatus for recommending content, a method and an apparatus for training a ranking model, a device, and a storage medium, which relate to a field of computer technology, in particular to fields of big data, deep learning, and intelligent search technology. A specific implementation solution includes: determining a scenario for recommendation; selecting target sub-models corresponding to the scenario for recommendation from at least two sub-models, where the at least two sub-models include sub-models of at least two dimensions, the sub-models of the at least two dimensions are trained for the at least two dimensions based on general feature data and feature data of the at least two dimensions, the general feature data and the feature data of the at least two dimensions are constructed based on behavior information, and the behavior information includes behavior information of at least two scenarios; and recommending content for the scenario for recommendation by using a combination of the target sub-models. A ranking module for each scenario is not required to be constructed in the present disclosure, so that the model training and the recommendation process is simplified.
-
8.发明公开
公开(公告)号：EP4134900A2
公开(公告)日：2023-02-15
申请号：EP22210615.5
申请日：2022-11-30
发明人： XIA, Xiaoling , XU, Wei , CAO, Chengtai , CHEN, Kunbin , HE, Bolei , HE, Wei
IPC分类号： G06Q30/0601 , G06Q30/0251
摘要： The present disclosure provides a method and an apparatus for recommending content, a method and an apparatus for training a ranking model, a device, and a storage medium, which relate to a field of computer technology, in particular to fields of big data, deep learning, and intelligent search technology. A specific implementation solution includes: determining a scenario for recommendation; selecting target sub-models corresponding to the scenario for recommendation from at least two sub-models, where the at least two sub-models include sub-models of at least two dimensions, the sub-models of the at least two dimensions are trained for the at least two dimensions based on general feature data and feature data of the at least two dimensions, the general feature data and the feature data of the at least two dimensions are constructed based on behavior information, and the behavior information includes behavior information of at least two scenarios; and recommending content for the scenario for recommendation by using a combination of the target sub-models. A ranking module for each scenario is not required to be constructed in the present disclosure, so that the model training and the recommendation process is simplified.
-
9.发明公开
公开(公告)号：EP3889830A1
公开(公告)日：2021-10-06
申请号：EP21150239.8
申请日：2021-01-05
发明人： NIU, Guocheng , HE, Bolei , XIAO, Xinyan
摘要： Embodiments of the present disclosure provide a cross-modality processing method, a cross-modality processing apparatus, an electronic device, and a computer storage medium, and relate to a field of natural language processing technologies. The method includes: obtaining (101) a sample set; generating (102) a plurality of training samples according to the sample set; adopting (103) the plurality of the training samples to train a semantic model, so that the semantic model learns semantic vectors containing combinations of the corpus and the images. Therefore, the training effect of the semantic model is improved, and in the related art, when multi-modal processing is performed, each modal is separately trained, and the semantic relations between different modals are isolated, thus the effect of the trained model is poor.
-
公开(公告)号：EP3846069A1
公开(公告)日：2021-07-07
申请号：EP20185932.9
申请日：2020-07-15
发明人： GAO, Can , LIU, Hao , HE, Bolei , XIAO, Xinyan , TIAN, Hao
摘要： The present disclosure provides a pre-training method for a sentiment analysis model and an electronic device, which relates to a field of artificial intelligence technologies. The method includes: based on a given seed sentiment dictionary, performing sentimental knowledge detection on a training corpus in a training corpus set, and determining a detection sentiment word and a detection word pair of the training corpus; according to preset mask processing rules, performing mask process on the training corpus to generate a masked corpus; performing encoding and decoding on the masked corpus by using a preset encoder and decoder to determine the detection sentiment word and the detection word pair of the training corpus; and updating the preset encoder and decoder according to a difference between prediction sentiment word and the detection sentiment word, and a difference between prediction word pair and the detection word pair.
-
-
-
-
-
-
-
-
-
搜索结果
10 条记录 (耗时 0.015 秒)
