WO2018000281A1 - Système et procédé d'apprentissage de représentation de portrait d'utilisateur à base de réseau neuronal profond - Google Patents
Système et procédé d'apprentissage de représentation de portrait d'utilisateur à base de réseau neuronal profond Download PDFInfo
- Publication number
- WO2018000281A1 WO2018000281A1 PCT/CN2016/087773 CN2016087773W WO2018000281A1 WO 2018000281 A1 WO2018000281 A1 WO 2018000281A1 CN 2016087773 W CN2016087773 W CN 2016087773W WO 2018000281 A1 WO2018000281 A1 WO 2018000281A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- user
- neural network
- deep neural
- user portrait
- statement
- Prior art date
Links
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 title claims description 14
- 230000006870 function Effects 0.000 claims abstract description 18
- 238000013135 deep learning Methods 0.000 claims abstract description 11
- 239000013598 vector Substances 0.000 claims abstract description 10
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 239000000284 extract Substances 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 244000145845 chattering Species 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000011273 social behavior Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Definitions
- the invention relates to big data, artificial intelligence and machine learning, in particular to a user image representation learning system and method based on deep neural network.
- social networks have grown rapidly and the number of users has exploded.
- people are more likely to use social networks as a public media platform to meet social needs and specific interests.
- the current social network products can not meet the demand very well, the information published by various users is mixed, users need to identify the information they are interested in. If the deep learning extraction feature is selected for the information trend and distribution characteristics in the specific domain of the social network, it is more favorable to obtain the information characteristics of the user.
- user portraits are mainly implemented by means of statistics, which often ignores some of the hidden information of the user.
- the feature can be abstracted and refined, the feature representation is more concise and more precise, and the deeper hidden information can be extracted.
- the invention discloses a user image representation learning system based on a deep neural network, which comprises: an intent recognition module for identifying a user's use function according to the received sentence; and a feature vector extraction module for inputting the sentence text through deep learning
- the context relationship is modeled, and the feature information extracted by the user is extracted; and the user portrait learning module is used to continuously update the user portrait through the iterative training of the feature information and the supervision information.
- the method further includes a statement receiving module, configured to receive a statement input by the user.
- a statement receiving module configured to receive a statement input by the user.
- the user-entered statement contains statements for unstructured data processing and statements that can be converted to structured data.
- the unstructured data refers to statement text input by a user.
- the structured data refers to data after converting an input sentence into an entity and a relationship.
- the function of identifying the user includes multiple functions.
- the invention also discloses a user image representation learning method based on deep neural network, wherein the package. Including: identifying the user's use function; extracting the feature information of the use function; and continuously updating the user's portrait through the iterative training of the feature information and the supervisory information.
- the statement further includes receiving a user input.
- the user-entered statement includes statements for unstructured data processing and statements that can be converted to structured data.
- the unstructured data refers to statement text input by a user.
- the structured data refers to data after converting an input sentence into an entity and a relationship.
- the supervisory information is combined when updating the user's portrait.
- FIG. 1 is a block diagram of a user image representation learning system based on a deep neural network according to an embodiment of the present invention
- FIG. 2 is a flow chart of a method for user portrait representation learning based on a deep neural network according to an embodiment of the present invention.
- the user portrait representation learning system 100 includes a sentence receiving module 101, an intent recognition module 102, a feature vector extraction module 103, and a user portrait learning module 104.
- the statement receiving module is configured to receive a statement input by the user, and send the statement to the user after receiving the statement of the user.
- the identification module 102 is configured to identify a function that the user wants to use according to a user's sentence, such as a function that the user wants to chat, order, or tune, and the like, and may be one or more of a plurality of functions.
- structured data refers to data after converting an input statement into an entity and a relationship.
- the chat input sentence can be used as unstructured data
- the on-demand input statement data can extract the corresponding on-demand entity.
- the statement entered by the user contains statements for unstructured data processing and statements that can be converted to structured data.
- the statement text entered by the user is set to unstructured data.
- the feature vector extraction module 103 models the context relationship of the input sentence text by deep learning according to the unstructured data, and extracts the feature information value of the user using the function.
- the user portrait learning module 104 continuously updates the user portrait by iteratively training the feature information and the supervisory information.
- structured data deep learning is used to model the relationship between entities, and then the user image is updated with a small amount of supervised information to finally obtain high-quality user images.
- structured data refers to data after converting an input statement into an entity and a relationship.
- step S201 the user inputs the sentence text; the user's intention is determined based on the sentence text for input.
- the intent information of the user is obtained, for example, the user inputs a chat statement, an on-demand statement, or a training statement.
- Feature extraction is performed on the obtained user intent information, mainly by extracting feature vectors through deep learning.
- step S203 extracting a feature vector of the user intention information, such as a chat party vector, an on-demand vector or a training vector, and the like.
- step S204 the user portrait is updated, and the user image is continuously updated through the iterative training of the feature information and the supervision information.
- the present invention is described in the following three intents as an embodiment, but it should be understood that the present invention is not limited to the input of the three intentions (i.e., chattering intention, on-demand intention, or tuning intention), or Other types of intent input.
- the statement text entered by the user is set to unstructured data.
- the context relationship of the input sentence text is modeled by deep learning, and the feature information of the function used by the user is extracted.
- the feature information value the user image is updated in combination with a small amount of supervision information.
- structured data use deep learning to model the relationship between entities, and then update the user's portrait with a small amount of supervised information to achieve high quality.
- User portrait refers to data after converting an input statement into an entity and a relationship.
- the present invention proposes a user image representation learning system and method for deep neural networks, which analyzes user characteristics and updates user images to obtain high quality user images.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Primary Health Care (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Machine Translation (AREA)
- Image Analysis (AREA)
Abstract
L'invention concerne un système d'apprentissage de représentation de portrait d'utilisateur à base de réseau neuronal profond, comprenant : un module de reconnaissance d'intention (102) utilisé pour reconnaître une fonction d'utilisation d'un utilisateur selon une déclaration reçue; un module d'extraction de vecteur caractéristique (103) utilisé pour modéliser la relation de contexte d'un texte ou la relation entre des entités grâce à un apprentissage profond, puis extraire des informations caractéristiques de l'utilisateur grâce à des informations textuelles saisies par l'utilisateur; et un module d'apprentissage de portrait d'utilisateur (104) utilisé pour mettre à jour en continu un portrait d'utilisateur grâce à un apprentissage itératif des informations caractéristiques et des informations de supervision. Grâce à l'apprentissage d'un portrait d'utilisateur dans un mode d'apprentissage profond, les caractéristiques du portrait d'utilisateur peuvent être extraites de manière abstraite, la représentation de caractéristique est plus concise et précise, et un niveau profond d'informations implicites peut être extrait.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2016/087773 WO2018000281A1 (fr) | 2016-06-29 | 2016-06-29 | Système et procédé d'apprentissage de représentation de portrait d'utilisateur à base de réseau neuronal profond |
CN201680001741.9A CN106489159A (zh) | 2016-06-29 | 2016-06-29 | 一种基于深度神经网络的用户画像表示学习系统及方法 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2016/087773 WO2018000281A1 (fr) | 2016-06-29 | 2016-06-29 | Système et procédé d'apprentissage de représentation de portrait d'utilisateur à base de réseau neuronal profond |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2018000281A1 true WO2018000281A1 (fr) | 2018-01-04 |
Family
ID=58286097
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2016/087773 WO2018000281A1 (fr) | 2016-06-29 | 2016-06-29 | Système et procédé d'apprentissage de représentation de portrait d'utilisateur à base de réseau neuronal profond |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN106489159A (fr) |
WO (1) | WO2018000281A1 (fr) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108828948A (zh) * | 2018-07-15 | 2018-11-16 | 大国创新智能科技(东莞)有限公司 | 基于深度学习的人工智能作战方法和机器人系统 |
CN109918162A (zh) * | 2019-02-28 | 2019-06-21 | 集智学园(北京)科技有限公司 | 一种可学习的海量信息高维图形交互式展示方法 |
CN112364008A (zh) * | 2020-11-20 | 2021-02-12 | 国网江苏省电力有限公司营销服务中心 | 一种面向电力物联网智能终端的设备画像构建方法 |
CN112908481A (zh) * | 2021-03-18 | 2021-06-04 | 马尚斌 | 一种自动化个人健康评估及管理方法及系统 |
CN113282757A (zh) * | 2021-07-14 | 2021-08-20 | 国网电子商务有限公司 | 基于电商领域表示模型的端到端三元组提取方法及系统 |
CN113609851A (zh) * | 2021-07-09 | 2021-11-05 | 浙江连信科技有限公司 | 心理学上想法认知偏差的识别方法、装置及电子设备 |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106934068A (zh) * | 2017-04-10 | 2017-07-07 | 江苏东方金钰智能机器人有限公司 | 机器人基于环境上下文的语义理解的方法 |
CN108804704A (zh) * | 2018-06-19 | 2018-11-13 | 北京顶象技术有限公司 | 一种用户深度画像方法及装置 |
CN109783733B (zh) * | 2019-01-15 | 2020-11-06 | 腾讯科技(深圳)有限公司 | 用户画像生成装置及方法、信息处理装置及存储介质 |
CN113326426A (zh) * | 2020-11-17 | 2021-08-31 | 崔海燕 | 基于大数据定位和人工智能的信息推送方法及系统 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104933049A (zh) * | 2014-03-17 | 2015-09-23 | 华为技术有限公司 | 生成数字人的方法及系统 |
US20150278688A1 (en) * | 2014-03-25 | 2015-10-01 | Nanyang Technological University | Episodic and semantic memory based remembrance agent modeling method and system for virtual companions |
CN105068661A (zh) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术(北京)有限公司 | 基于人工智能的人机交互方法和系统 |
CN105096170A (zh) * | 2015-09-18 | 2015-11-25 | 车智互联(北京)科技有限公司 | 基于bbd或/和rf模型获取潜客级别的方法和系统 |
CN105183848A (zh) * | 2015-09-07 | 2015-12-23 | 百度在线网络技术(北京)有限公司 | 基于人工智能的人机聊天方法和装置 |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101076061A (zh) * | 2007-03-30 | 2007-11-21 | 腾讯科技(深圳)有限公司 | 一种机器人服务器及自动聊天方法 |
CN104750731B (zh) * | 2013-12-30 | 2018-09-21 | 华为技术有限公司 | 一种获取完整用户画像的方法及装置 |
CN105446973B (zh) * | 2014-06-20 | 2019-02-26 | 华为技术有限公司 | 社交网络中用户推荐模型的建立及应用方法和装置 |
CN104881594B (zh) * | 2015-05-06 | 2018-04-03 | 镇江乐游网络科技有限公司 | 一种基于精准画像的智能手机拥有权检测方法 |
CN104850662B (zh) * | 2015-06-08 | 2018-05-29 | 浙江每日互动网络科技股份有限公司 | 一种基于用户画像的移动终端智能消息推送方法、服务器和系统 |
CN105005587A (zh) * | 2015-06-26 | 2015-10-28 | 深圳市腾讯计算机系统有限公司 | 一种用户画像的更新方法、装置和系统 |
CN105574159B (zh) * | 2015-12-16 | 2019-04-16 | 浙江汉鼎宇佑金融服务有限公司 | 一种基于大数据的用户画像建立方法和用户画像管理系统 |
CN105589956B (zh) * | 2015-12-21 | 2018-11-27 | 东软集团股份有限公司 | 一种用户画像的方法及装置 |
CN105608171B (zh) * | 2015-12-22 | 2018-12-11 | 青岛海贝易通信息技术有限公司 | 用户画像构建方法 |
-
2016
- 2016-06-29 WO PCT/CN2016/087773 patent/WO2018000281A1/fr active Application Filing
- 2016-06-29 CN CN201680001741.9A patent/CN106489159A/zh active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104933049A (zh) * | 2014-03-17 | 2015-09-23 | 华为技术有限公司 | 生成数字人的方法及系统 |
US20150278688A1 (en) * | 2014-03-25 | 2015-10-01 | Nanyang Technological University | Episodic and semantic memory based remembrance agent modeling method and system for virtual companions |
CN105068661A (zh) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术(北京)有限公司 | 基于人工智能的人机交互方法和系统 |
CN105183848A (zh) * | 2015-09-07 | 2015-12-23 | 百度在线网络技术(北京)有限公司 | 基于人工智能的人机聊天方法和装置 |
CN105096170A (zh) * | 2015-09-18 | 2015-11-25 | 车智互联(北京)科技有限公司 | 基于bbd或/和rf模型获取潜客级别的方法和系统 |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108828948A (zh) * | 2018-07-15 | 2018-11-16 | 大国创新智能科技(东莞)有限公司 | 基于深度学习的人工智能作战方法和机器人系统 |
CN109918162A (zh) * | 2019-02-28 | 2019-06-21 | 集智学园(北京)科技有限公司 | 一种可学习的海量信息高维图形交互式展示方法 |
CN109918162B (zh) * | 2019-02-28 | 2021-11-02 | 集智学园(北京)科技有限公司 | 一种可学习的海量信息高维图形交互式展示方法 |
CN112364008A (zh) * | 2020-11-20 | 2021-02-12 | 国网江苏省电力有限公司营销服务中心 | 一种面向电力物联网智能终端的设备画像构建方法 |
CN112908481A (zh) * | 2021-03-18 | 2021-06-04 | 马尚斌 | 一种自动化个人健康评估及管理方法及系统 |
CN112908481B (zh) * | 2021-03-18 | 2024-04-16 | 马尚斌 | 一种自动化个人健康评估及管理方法及系统 |
CN113609851A (zh) * | 2021-07-09 | 2021-11-05 | 浙江连信科技有限公司 | 心理学上想法认知偏差的识别方法、装置及电子设备 |
CN113282757A (zh) * | 2021-07-14 | 2021-08-20 | 国网电子商务有限公司 | 基于电商领域表示模型的端到端三元组提取方法及系统 |
Also Published As
Publication number | Publication date |
---|---|
CN106489159A (zh) | 2017-03-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2018000281A1 (fr) | Système et procédé d'apprentissage de représentation de portrait d'utilisateur à base de réseau neuronal profond | |
CN107886949B (zh) | 一种内容推荐方法及装置 | |
US10949748B2 (en) | Deep learning of bots through examples and experience | |
WO2020082560A1 (fr) | Procédé, appareil et dispositif d'extraction de mot-clé de texte, ainsi que support de stockage lisible par ordinateur | |
US20160328469A1 (en) | Method, Device and Equipment for Acquiring Answer Information | |
US20180203847A1 (en) | Tone optimization for digital content | |
US9536444B2 (en) | Evaluating expert opinions in a question and answer system | |
CN103399906B (zh) | 在进行输入时基于社会关系提供候选词的方法和装置 | |
CN106934068A (zh) | 机器人基于环境上下文的语义理解的方法 | |
CN114303132A (zh) | 在虚拟个人助手中使用唤醒词进行上下文关联和个性化的方法和系统 | |
US11380131B2 (en) | Method and device for face recognition, storage medium, and electronic device | |
CN111160452A (zh) | 一种基于预训练语言模型的多模态网络谣言检测方法 | |
CN111400504A (zh) | 企业关键人的识别方法和装置 | |
WO2018068648A1 (fr) | Procédé de mise en correspondance d'informations et dispositif connexe | |
WO2020103606A1 (fr) | Procédé et dispositif de traitement de modèle, terminal et support de stockage | |
CN111368145A (zh) | 一种知识图谱的创建方法、创建系统及终端设备 | |
CN107609487B (zh) | 一种用户头像的生成方法和装置 | |
CN108306813B (zh) | 会话消息的处理方法、服务器及客户端 | |
US10217455B2 (en) | Linguistic model database for linguistic recognition, linguistic recognition device and linguistic recognition method, and linguistic recognition system | |
CN114564595A (zh) | 知识图谱更新方法、装置及电子设备 | |
CN109190116B (zh) | 语义解析方法、系统、电子设备及存储介质 | |
CN113220828A (zh) | 意图识别模型处理方法、装置、计算机设备及存储介质 | |
CN112269928A (zh) | 用户推荐方法及装置、电子设备、计算机可读介质 | |
CN105376143B (zh) | 一种识别发送者身份的方法及装置 | |
CN109961152B (zh) | 虚拟偶像的个性化互动方法、系统、终端设备及存储介质 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 16906682 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 16906682 Country of ref document: EP Kind code of ref document: A1 |