WO2020238498A1 - 问答信息的处理方法、系统、计算机设备和存储介质 - Google Patents

问答信息的处理方法、系统、计算机设备和存储介质 Download PDF

Info

Publication number
WO2020238498A1
WO2020238498A1 PCT/CN2020/086128 CN2020086128W WO2020238498A1 WO 2020238498 A1 WO2020238498 A1 WO 2020238498A1 CN 2020086128 W CN2020086128 W CN 2020086128W WO 2020238498 A1 WO2020238498 A1 WO 2020238498A1
Authority
WO
WIPO (PCT)
Prior art keywords
mark
page
question
answer
information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/086128
Other languages
English (en)
French (fr)
Inventor
郭鸿程
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
OneConnect Smart Technology Co Ltd
Original Assignee
OneConnect Smart Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by OneConnect Smart Technology Co Ltd filed Critical OneConnect Smart Technology Co Ltd
Publication of WO2020238498A1 publication Critical patent/WO2020238498A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33Querying
    • G06F16/332Query formulation
    • G06F16/3329Natural language query formulation
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44Arrangements for executing specific programs
    • G06F9/445Program loading or initiating
    • G06F9/44521Dynamic linking or loading; Link editing at or after load time, e.g. Java class loading
    • G06F9/44526Plug-ins; Add-ons
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44Arrangements for executing specific programs
    • G06F9/451Execution arrangements for user interfaces

Definitions

  • This application relates to the field of artificial intelligence technology, and specifically to a method for processing question and answer information, a system for processing question and answer information, computer equipment and storage media.
  • the purpose of this application is to solve at least one of the above technical defects, especially the technical defect of low efficiency of the problem process.
  • This application provides a method for processing question and answer information, including the following steps:
  • mark input control Setting a mark input control on the page of the product display, wherein the mark input control is used to record and display the mark traces generated on the page;
  • the question information is input into a question and answer model, and a target answer is obtained and output on the page displayed on the product, where the question and answer model is a model for answering questions about the products displayed on the page.
  • This application also provides a system for processing question and answer information, including:
  • a setting module configured to set a mark input control on the product display page, wherein the mark input control is used to record and display the mark traces generated on the page;
  • An acquiring module configured to acquire the mark input by the user on the page when the mark input control is triggered, and record the trace and position of the mark;
  • An identification module configured to identify the user’s problem information according to the trace and location of the mark and the page displayed on the product
  • the output module is used to input the question information into a question and answer model, and obtain and output a target answer on the page displayed on the product, wherein the question and answer model is a model used to answer questions about the product displayed on the page.
  • the present application also provides a computer device, including a memory and a processor, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes any implementation The steps of the method of processing question and answer information described in the example.
  • the present application also provides a storage medium storing computer-readable instructions.
  • the computer-readable instructions are executed by one or more processors, the one or more processors execute the processing of the question and answer information in any embodiment. Method steps.
  • the above-mentioned question and answer information processing method, system, computer equipment and storage medium can conveniently and easily input the mark for asking the question on the current product display page, and identify the question raised according to the mark and the content information of the page Information simplifies the process of asking questions and can improve the efficiency of the question and answer process. Moreover, by inputting question information into the intelligent question-and-answer model with reading comprehension function, the target answer closely related to the content information of the page can be obtained, and the accuracy of the target answer can be improved.
  • Figure 1 is an implementation environment diagram of a method for processing question and answer information provided in an embodiment
  • Figure 2 is a flowchart of a method for processing question and answer information in an embodiment
  • Figure 3 is a flow chart of identifying problem information in an embodiment
  • Figure 4 is a page diagram of a marked product display in an embodiment
  • Figure 5 is a flowchart of a method for processing question and answer information in another embodiment
  • Fig. 6 is a schematic structural diagram of a system for processing question and answer information in an embodiment
  • Fig. 7 is a schematic diagram of the internal structure of a computer device in an embodiment.
  • FIG. 1 is an implementation environment diagram of a method for processing question and answer information provided in an embodiment.
  • the implementation environment includes a server device 110 and a terminal 120.
  • the server device 110 may be a computer, an independent server, or a server cluster composed of multiple servers.
  • the server device 110 or the terminal 120 may be installed with a question and answer model.
  • the question and answer model may have the semantics of recognizing the read text and combine the read text with the input The ability to contact questions and output corresponding answers.
  • the terminal 120 has a data input function.
  • the terminal 120 can be connected to a mouse and record the mouse track, and the terminal 120 can also be equipped with a touch screen, which can record the user's touch track.
  • the terminal 120 can be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc., but is not limited thereto.
  • the server device 110 and the terminal 120 can communicate through a network.
  • FIG. 2 is a flowchart of a method for processing question and answer information in an embodiment.
  • a method for processing question and answer information is proposed.
  • the method for processing question and answer information can be applied In the above implementation environment, the following steps can be specifically included:
  • Step S210 Set a mark input control on the product display page, where the mark input control is used to record and display the mark traces generated on the page.
  • a mark input control is set on the product display page.
  • the mark input control can be directly triggered on the current page without jumping to other pages to find the mark input Control.
  • Step S220 When the mark input control is triggered, obtain the mark input by the user on the page, and record the trace and position of the mark.
  • the traces of the user's touch movement or the movement of the mouse on the page where the product is displayed can quickly produce marks on the page.
  • the traces of the user's touch and movement or the traces of the mouse movement in the page of browsing the product display can be recorded, and these movement traces are used as the input mark.
  • Entering tags on the product display page can avoid the need for users to go back and forth on the product display page and the page of the customer service dialog when asking questions, reducing the process of page round trips; it can also reduce the user organization and input text information of the problem Time, improve the efficiency of questioning.
  • the mark input by the user at a certain time can be a combination of several discrete marks, and each mark in the mark and the position of each mark can be recorded. For some traces that are relatively concentrated or have interlaced tracks, these traces can be combined and the overall position of the traces can be recorded. In addition, the position of the entire area occupied by the mark can be recorded. At this time, the area where the entire mark is located can match the part of the display page, and the subsequent identified problem information can be associated with the part of the display page to improve the identification The accuracy of the problem information.
  • the mark may include a touch track and/or a mouse track.
  • the terminal 120 used by the user to browse the product display page is a display device with touch screen function
  • the touch on the touch screen can be sensed, and the trace of the touch can be recorded at the same time, that is, the touch track is recorded.
  • the touch trace serves as a mark, and the position of the mark relative to the page displayed on the product is identified.
  • the browsing terminal 120 triggers the mark input control and the input mark by clicking the mouse, the mark is obtained by recording the trace of the mouse click, that is, the mark in the product display page
  • the mouse track serves as the mark input by the user, and also identifies the position of the mark relative to the page displayed on the product.
  • the touch track and the mouse track can be recorded at the same time as the mark input by the user on the page.
  • the touch track and/or mouse track can be selected adaptively as the input mark according to the device with which the user browses the page.
  • the Q&A information processing method can use touch track or mouse track as the input mark according to the user's equipment capabilities.
  • This mark input method is highly interactive. It can be easily and freely entered and entered through touch contact or mouse operation.
  • the mark related to the question information, for users who are not familiar with the product display page, can also quickly grasp how to enter the mark for questioning.
  • Step S230 Identify the user's problem information according to the marked trace and location and the page displayed on the product.
  • the trace of the mark is extracted and the location of the trace on the product display page is identified.
  • the obtained mark is used to identify the problem information expressed by the user, and the trace and location of the recorded mark can indirectly describe what is involved in the problem information Text information.
  • the marked traces can display the text information expressed by the user through text handwriting, symbol handwriting and marking graphics; in addition, the combination of the marked position and the content information on the product display page can also identify the text information referred to by the mark; Alternatively, the semantics of the recognized text information can also be analyzed through the position of the mark.
  • Step S240 Input the question information into the question and answer model, and obtain and output the target answer on the product display page, where the question and answer model is a model for answering questions about the product displayed on the page.
  • the question and answer model has an intelligent reading comprehension function.
  • the question and answer model can interpret the content information displayed on the page, and combine and analyze the question information with the content information displayed on the page , Get the target answer used to answer the question information, that is, get the target answer.
  • the target answer can also be output in the page where the product is displayed.
  • the target page can be output in a partial area of the page.
  • the target answer and the content information of the product display can be compared and read on the same page to meet the requirements of reading the product display.
  • the page allows users to receive the target answer, which is convenient for users to quickly read and understand the target answer and the content information of the product display, and improve reading efficiency.
  • the above method for processing question and answer information can conveniently and easily input the mark for asking the question on the current product display page, and identify the question information raised according to the mark and the content information of the page, simplifying the process of raising the question information , Can improve the efficiency of the question and answer process. Moreover, by inputting question information into the intelligent question-and-answer model with reading comprehension function, the target answer closely related to the content information of the page can be obtained, and the accuracy of the target answer can be improved.
  • the identified question information may be displayed in a partial area of the page, so as to facilitate understanding of the target answer based on the question information.
  • the output form of the target answer can be determined according to a user-selected or preset answer output model.
  • the step of outputting the target answer on the page of the product display in step S240 may include:
  • first search for the preset answer output mode determine the information format selected by the user or corresponding to the preset answer output mode, and output the target answer in the information format.
  • the following will convert the target answer according to the information format corresponding to the answer output mode.
  • the target answer in the voice output mode, can be broadcast by voice, and the user can understand the target answer by browsing the page while listening, staying on the page of the product currently viewed, and reducing page jumps.
  • a voice control used to trigger the playback of the voice signal corresponding to the target answer can be set in a certain position of the page setting of the product display.
  • the voice control can also be used to collect voice signals input by the user. After obtaining the input voice signal, obtain the question information contained in the input voice signal through voice recognition, and execute the steps of inputting the question information into the question and answer model, obtaining and outputting the target answer on the product display page, and completing the voice input To ask questions and answer questions.
  • the target answer in the text output mode, can be displayed through text information, and a text display window is set on a part of the product display page.
  • the text display window can be used to display the target answer and display the content of the product display.
  • Information, the target answer and product display content information can be displayed on the same page, reducing page jumps.
  • the above question and answer information processing method can provide a variety of answer output modes, and output the target answer in the corresponding information format according to the preset answer output mode, adapting to the user's reading habits; it can also reduce or avoid page stress when outputting the target answer Jump, easy to understand the target answer combined with the content information of the product display.
  • FIG. 3 is a flowchart of identifying problem information in an embodiment.
  • the step of identifying the user's problem information according to the marked trace and location and the page displayed on the product in step S230 may include :
  • Step S231 Divide the marking information into several marks according to the marks and positions of the marks.
  • a certain output mark can contain several discontinuous drag or touch traces, but there are correlations between these traces.
  • the combination of traces in a local area can show the shape of text or symbols, and some traces are still Can show graphics of written notes.
  • a group of related traces can be divided into a mark.
  • the relationship or characteristics between the traces can be judged by the traces and positions of the marks.
  • Different types of marks have specific characteristics. Taking text marks as an example, the positions of the traces that can constitute the text will be relatively close, and there will often be overlaps between the traces; for example, underline, the length of the traces will be relatively long; Take the circle mark as an example, the curvature and curvature of the trace will be larger. Then, according to the correlation between the traces, the traces in the mark are divided into several marks.
  • Step S232 Identify the symbol type of each symbol.
  • the mark type of the mark can be judged and the mark type of each mark can be identified based on the characteristics of the marks in the mark or the association relationship between the marks.
  • Step S233 According to the product display page and the mark type of each mark, the text information corresponding to each mark is identified one by one.
  • Each mark will contain the referenced text information, and the mark type of each mark further identifies the text it refers to. Since the mark is generated on the page of the product display, some marks need to be combined with the content information in the page to find the text information referred to. Therefore, the text information corresponding to each mark needs to be identified in combination with the page of the product display.
  • the step of identifying the text information corresponding to each mark one by one may include:
  • A1 When the mark type of the mark is a text track, convert the mark into a handwriting image, recognize the text information in the handwriting image, and use the text information in the handwriting image as the text information corresponding to the mark.
  • the marks of the mark can be superimposed together to visually observe the corresponding character or symbol of the mark, and then use image processing to recognize the character or symbol, and use the recognized character or symbol as the corresponding mark Text message.
  • A2 When the type of the mark is underline or circle mark, according to the position of the mark on the product display page, extract the text information corresponding to the position from the text content of the product display, and use the text information corresponding to the position as the text information corresponding to the mark .
  • the graphic mark may include an underline or a circle, and the text information displayed at the position of the graphic mark is searched according to the graphic mark.
  • the line where the text information is located can be determined according to the position of the underline, the starting position of the text information can be determined from the line according to the first and last positions of the underline, and the text information can be extracted from the content information on the product display page.
  • Text information; in addition, the text information covered by the circle can also be determined according to the position of the circle, and the covered text information can be extracted from the content information on the product display page.
  • the product display image at the location of the graphic mark can also be intercepted on the product display page, and the words and symbols in the intercepted product display image can be recognized through image processing, and the recognized words and symbols can be used as Text information corresponding to the mark.
  • the above-mentioned implementation of identifying the text information corresponding to each mark can effectively and accurately convert the mark into text information; the obtained text information can be recognized by the computer and the question-and-answer model, so that the question-and-answer model can be called subsequently, and the efficiency and efficiency of question answering can be improved. Accuracy.
  • mark 1, mark 2, mark 3, and mark 4 are used as the mark of the product display page.
  • mark 1, mark 2, and mark 4 the marks are converted into images, and image processing is performed to identify the characters " ⁇ ", the characters "Long” and the symbol "?”.
  • the angle of mark 3 deviating from the horizontal line is low, and the type of mark 3 can be determined to be underlined; get the position of mark 3 relative to the product display page and the first and last positions of mark 3 from the page
  • the content information extracts the text information "for those that occurred before the expiration of the main insurance contract after the waiting period and continued within 30 days after the expiration of the main insurance contract".
  • Step S234 Generate question information according to the text information.
  • the text information is combined and processed according to the grammatical rules of the natural language to generate question information that conforms to the semantics of the natural language.
  • the step of generating question information based on text information may include:
  • B2 Get the position of the mark between the text, punctuation mark and mark.
  • the question information expressed by the user can be artificially recognized, and the efficiency and intelligence of question information input can be improved.
  • the page of the product display is used to display insurance clauses
  • the question-and-answer model is used to answer the question information under the requirements of the insurance clauses.
  • the product display page in Figure 4 Take the marked product display page in Figure 4 as an example.
  • the product display page displays the insurance clauses, but the insurance clauses are very professional and often difficult to understand when reading. You must clearly express the questions you want to ask. It is also very difficult.
  • the page of the product display in Figure 4 contains multiple terms in "for those that occur before the expiration of the main insurance contract after the waiting period and continue within 30 days after the expiration of the main insurance contract", the user should You need to ask questions about the terms, you need to repeat the terms, back and forth between the terms page and the customer service dialog box. This makes it difficult for users who do not understand the terms to describe the problem when they come, and greatly complicates the process of user questions.
  • the clause when the clause is marked, the clause can be automatically identified according to the mark and the customer's question about the clause can be generated, which simplifies the process for the user to ask questions.
  • the question-and-answer model can have natural language reading and comprehension ability, and after analyzing the actual semantics of the insurance product's clauses, it can answer the question information.
  • the output target answer is strongly related to the insurance clause, and the target answer is accurate high.
  • the above-mentioned method of processing question and answer information can improve the accuracy of answering questions about insurance regulations and improve the quality of customer service of insurance products.
  • FIG. 5 is a flowchart of a method for processing question and answer information in another embodiment.
  • This embodiment proposes a method for processing question and answer information, which may specifically include the following steps:
  • Step S510 Set a mark input control on the insurance policy page of the insurance product.
  • the mark input control is used to record and display the touch track and/or the mouse track generated on the insurance policy page.
  • Step S520 When the mark input control is triggered, obtain the touch track and/or mouse track input by the user on the insurance policy page, and record the trace and position of the touch track and/or mouse track.
  • Step S530 Identify the user's insurance product problem based on the trace and location of the touch track and/or mouse track and the insurance policy page.
  • the touch track and/or mouse track are divided into several marks according to their traces and positions. Identify the symbol type of each symbol.
  • the text information corresponding to each symbol is identified one by one.
  • the mark type of the mark is a text track
  • the mark is converted into a handwriting image
  • the text information in the handwriting image is recognized
  • the text information in the handwriting image is used as the text information corresponding to the mark;
  • the type of the mark is underline or circle mark
  • the text information corresponding to the position is extracted from the text content of the insurance regulations, and the text information corresponding to the position is taken as the text information corresponding to the mark.
  • Generate insurance product questions based on text information Recognize the words and punctuation marks of text information; obtain the mark positions between words, punctuation marks and marks; generate insurance product questions based on the words, punctuation marks and mark positions, combined with semantics.
  • Step S540 The Q&A model is used to answer the insurance product questions in combination with the currently browsed insurance regulations, input the insurance product questions into the Q&A model, and obtain and output the target answers on the insurance regulations page.
  • the step of outputting the target answer on the insurance policy page may specifically include: obtaining the preset answer output mode; when the answer output mode is voice output, returning the voice signal corresponding to the target answer to the insurance policy page; and text output in the answer output mode At the time, the text information corresponding to the target answer is displayed on the insurance policy page.
  • the above-mentioned question and answer information processing method can conveniently and easily enter the mark reflecting the insurance product problem in the insurance regulations page, and identify the proposed insurance product problem according to the content information of the mark and insurance regulations, and it is easy to raise the insurance product problem
  • the process can improve the efficiency of the question and answer process of insurance product questions.
  • the question-and-answer model has the function of reading and understanding insurance regulations, and inputting insurance product questions into the question-and-answer model can obtain target answers that are closely related to the content information of the insurance regulations and improve the accuracy of the target answers.
  • FIG. 6 is a schematic structural diagram of a question and answer information processing system in an embodiment.
  • a question and answer information processing system is proposed.
  • the question and answer information processing system may specifically It includes a setting module 610, an acquisition module 620, an identification module 630, and an output module 640, where:
  • the setting module 610 is used to set a mark input control on the product display page, where the mark input control is used to record and display the mark traces generated on the page.
  • the setting module 610 sets a mark input control on the product display page.
  • the mark input control can be directly triggered on the current page without jumping to other pages to find the mark input control .
  • the obtaining module 620 is configured to obtain the mark input by the user on the page when the mark input control is triggered, and record the trace and position of the mark.
  • the traces of the user's touch movement or the movement of the mouse on the page where the product is displayed can quickly produce marks on the page.
  • the acquisition module 620 can record the traces of the user's touch and movement or the traces of the mouse movement in the page of browsing the product display, and use these movement traces as the input mark.
  • the identification module 630 is used to identify the user's problem information based on the marked trace and location and the page on which the product is displayed.
  • the recognition module 630 extracts the trace of the mark and recognizes the location of the trace on the product display page.
  • the obtained mark is used to identify the question information expressed by the user, and the recorded mark and location can indirectly describe the text involved in the question information information.
  • the output module 640 is used to input question information into the question and answer model, and obtain and output the target answer on the product display page, where the question and answer model is a model for answering questions about the product displayed on the page.
  • the output module 640 inputs the identified question information into the question-and-answer model.
  • the question-and-answer model has an intelligent reading comprehension function.
  • the question-and-answer model can interpret the content information displayed on the page and associate and analyze the question information in combination with the content information displayed on the page. Get the target answer used to answer the question information, that is, get the target answer.
  • the above-mentioned question and answer information processing system can conveniently and easily input the mark for asking questions on the current product display page, and identify the question information raised according to the mark and the content information of the page, simplifying the process of raising the question information , Can improve the efficiency of the question and answer process. Moreover, by inputting question information into an intelligent question-and-answer model with reading comprehension function, a target answer closely related to the content information of the page can be obtained, and the accuracy of the target answer can be improved.
  • Each module in the above-mentioned question and answer information processing system can be implemented in whole or in part by software, hardware, and a combination thereof.
  • the foregoing modules may be embedded in the form of hardware or independent of the processor in the computer device, or may be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the foregoing modules.
  • FIG. 7 is a schematic diagram of the internal structure of a computer device in an embodiment.
  • the computer device includes a processor, a nonvolatile storage medium, a memory, and a network interface connected through a system bus.
  • the non-volatile storage medium of the computer device stores an operating system, a database, and computer-readable instructions.
  • the database may store control information sequences.
  • the processor can realize a A method of processing question and answer information.
  • the processor of the computer equipment is used to provide calculation and control capabilities, and supports the operation of the entire computer equipment.
  • a computer readable instruction may be stored in the memory of the computer device.
  • the processor may execute a method for processing question and answer information.
  • the network interface of the computer device is used to connect and communicate with the terminal.
  • a computer device in one embodiment, includes a memory, a processor, and a computer program stored on the memory and running on the processor.
  • the processor executes the computer program to implement the question and answer in any of the above embodiments. Information processing method steps.
  • a storage medium storing computer-readable instructions.
  • the storage medium may be non-volatile or volatile.
  • the computer-readable instructions are executed by one or more processors. When executed, one or more processors are caused to execute the steps of the method for processing question and answer information in any of the foregoing embodiments.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

一种问答信息的处理方法、系统、计算机设备和存储介质,在产品展示的页面上设置标记输入控件,其中,标记输入控件用于记录并显示在页面上产生的标记的痕迹(S210);当标记输入控件被触发时,获取用户在页面上所输入的标记,并记录标记的痕迹和位置(S220);根据标记的痕迹和位置以及产品展示的页面识别用户的问题信息(S230);将问题信息输入问答模型,获取并在产品展示的页面输出目标答案(S240)。上述方法可以方便和简易地在当前的产品展示的页面上输入用于提问问题的标记,根据标记和页面的内容信息来识别所提出的问题信息,简便了提出问题信息的过程,提高问答过程的效率;通过向具有阅读理解功能的智能的问答模型输入问题信息,获得与页面的内容信息关系密切的目标答案,提高目标答案准确度。

Description

问答信息的处理方法、系统、计算机设备和存储介质
本申请要求于2019年5月24日提交中国专利局、申请号为201910440492.9,发明名称为“问答信息的处理方法、系统、计算机设备和存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能技术领域,具体而言,涉及一种问答信息的处理方法、问答信息的处理系统、计算机设备和存储介质。
背景技术
随着人们生活质量的提高以及社会水平的进步,大家对产品品质的关注度也越来越重视,对产品的介绍和展示内容认真细致地阅读。发明人意识到,有时候产品的内容信息是专业性较强的阅读内容,往往在浏览产品时难免会因难以理解而提出问题。特别是越来越受到公众的欢迎的保险产品,保险产品中的保险条例的专业性很强,保险条例在阅读上容易出现较多问题。
而一般在阅读展示产品的页面时,想到了疑问后需在跳转的客服对话框的页面中输入问题,但常常因操作过多,而忘记了要提问的问题如何表述,导致为了提问而来回在多个页面折返,非常耗费时间,导致提出问题信息的过程复杂,导致问答过程的效率低。
发明内容
本申请的目的旨在至少能解决上述的技术缺陷之一,特别是问题过程效率低的技术缺陷。
本申请提供一种问答信息的处理方法,包括如下步骤:
在产品展示的页面上设置标记输入控件,其中,所述标记输入控件用于记录并显示在所述页面上产生的标记的痕迹;
当所述标记输入控件被触发时,获取用户在所述页面上所输入的标记,并记录所述标记的痕迹和位置;
根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息;
将所述问题信息输入问答模型,获取并在所述产品展示的页面输出目标答案,其中,所述问答模型为用于解答所述页面所展示产品的问题的模型。
本申请还提供一种问答信息的处理系统,包括:
设置模块,用于在产品展示的页面上设置标记输入控件,其中,所述标记输入控件用于记录并显示在所述页面上产生的标记的痕迹;
获取模块,用于当所述标记输入控件被触发时,获取用户在所述页面上所输入的标记,并记录所述标记的痕迹和位置;
识别模块,用于根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息;
输出模块,用于将所述问题信息输入问答模型,获取并在所述产品展示的页面输出目标答案,其中,所述问答模型为用于解答所述页面所展示产品的问题的模型。
本申请还提供一种计算机设备,包括存储器和处理器,所述存储器中存储有计算机可读指令,所述计算机可读指令被所述处理器执行时,使得所述处理器执行如任一实施例所述问答信息的处理方法的步骤。
本申请还提供一种存储有计算机可读指令的存储介质,所述计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行任一实施例所述问答信息的处理方法的步骤。
上述的问答信息的处理方法、系统、计算机设备和存储介质,可以方便和简易地在当前的产品展示的页面上输入用于提问问题的标记,根据标记和页面的内容信息来识别所提出的问题信息,简便了提出问题信息的过程,可以提高问答过程的效率。而且通过向具有阅读理解功能的智能的问答模型输入问题信息,可以获得与页面的内容信息关系密切的目标答案,提高目标答案的准确度。
附图说明
本申请上述的和/或附加的方面和优点从下面结合附图对实施例的描述中将变得明显和容易理解,其中:
图1为一个实施例中提供的问答信息的处理方法的实施环境图;
图2为一个实施例中问答信息的处理方法的流程图;
图3为一个实施例中识别问题信息的流程图;
图4为一个实施例中带标记的产品展示的页面图;
图5为另一个实施例中问答信息的处理方法的流程图;
图6为一个实施例中问答信息的处理系统的结构示意图;
图7为一个实施例中计算机设备的内部结构示意图。
具体实施方式
下面详细描述本申请的实施例,所述实施例的示例在附图中示出,其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的,仅用于解释本申请,而不能解释为对本申请的限制。
本技术领域技术人员可以理解,除非另外定义,这里使用的所有术语(包括技术术语和科学术语),具有与本申请所属领域中的普通技术人员的一般理解相同的意义。还应该理解的是,诸如通用字典中定义的那些术语,应该被理解为具有与现有技术的上下文中的意义一致的意义,并且除非像这里一样被特定定义,否则不会用理想化或过于正式的含义来解释。
如图1所示,图1为一个实施例中提供的问答信息的处理方法的实施环境图,在该实施环境中,包括服务器设备110以及终端120。
服务器设备110可以是计算机、独立的服务器或者是多个服务器组成的服务器集群,服务器设备110或终端120上可以安装有问答模型,该问答模型可以具有识别阅读文本的语义、将阅读文本与所输入问题进行联系并输出对应答案的能力。终端120上具有数据输入的功能,例如终端120可以连接鼠标并记录鼠标轨迹,终端120也可以安装有触摸屏,该触摸屏可以记录用户的触摸轨迹。终端120可为智能手机、平板电脑、笔记本电脑、台式计算机等,但并不局限于此。服务器设备110以及终端120可以通过网络进行通信。
在一个实施例中,如图2所示,图2为一个实施例中问答信息的处理方法的流程图,本实施例中提出了一种问答信息的处理方法,该问答信息的处理方法可以应用于上述的实施环境中,具体可以包括以下步骤:
步骤S210:在产品展示的页面上设置标记输入控件,其中,标记输入控件用于记录并显示在页面上产生的标记的痕迹。
本步骤中,在产品展示的页面上设置有标记输入控制,在浏览当前的产品展示的页面的过程中,可以在当前页面直接触发该标记输入控件,无需跳转至其他页面来查找该标记输入控件。
步骤S220:当标记输入控件被触发时,获取用户在页面上所输入的标记,并记录标记的痕迹和位置。
用户在浏览产品展示的页面中触摸移动的痕迹或鼠标移动的痕迹急可以在页面上产生标记的痕迹。本步骤中,控件触发后,可以记录用户在浏览产品展示的页面中触摸移动的痕迹或鼠标移动的痕迹,将这些移动的痕迹作为所输入的标记。
在产品展示的页面上输入标记,可以避免用户在提出提问问题时需要在产品展示页面和客服对话框的页面中折返查看,减少页面往返的过程;还可以减少用户组织和输入问题的文本信息的时间,提高提问的效率。
用户在某次输入的标记可以是由若干个不连续的痕迹组合而成的,可以记录标记中各个痕迹以及各个痕迹的位置。对于部分在位置较为集中或轨迹具有交错的部分痕迹,可以将这部分痕迹进行组合,并记录该部分痕迹的整体位置。另外还可以记录该标记整体所占区域的位置,此时标记整体所在区域可以与展示页面中的局部匹配,后续所识别的问题信息可以与该展示页面中的局部的内容进行关联,提高所识别的问题信息的准确度。
进一步地,标记可以包括触摸轨迹和/或鼠标轨迹。当用户浏览产品 展示的页面所用的终端120是具有触摸屏功能的显示设备时,在标记输入控件启用过程中,可以感应触摸显示屏上的触摸,同时记录该触摸的痕迹,即记录触摸轨迹,将该触摸的痕迹作为标记,以及识别该标记相对于产品显示的页面的位置。另外,当用户浏览产品展示的页面时,如果浏览的终端120是通过鼠标点击的方式来触发标记输入控件和输入标记的,则通过记录鼠标点击的痕迹来获得标记,即将产品展示的页面中的鼠标轨迹作为用户所输入的标记,同时也识别该标记相对于产品显示的页面的位置。对于一些兼备外置鼠标输入和触摸屏功能的终端120,可以同时将触摸轨迹和鼠标轨迹进行记录,作为用户在页面上所输入的标记。可以根据用户浏览页面的设备,适应性地选择触摸轨迹和/或鼠标轨迹作为输入的标记。
此时,问答信息的处理方法可以根据用户的设备能力,使用触摸轨迹或鼠标轨迹作为所输入的标记,这种标记输入方式的交互性强,通过触摸接触或鼠标操作可以便捷且随心地输入与提问信息相关的标记,对于不熟悉产品展示页面的用户,也能够快速掌握如何输入用于提问的标记。
步骤S230:根据标记的痕迹和位置以及产品展示的页面识别用户的问题信息。
本步骤中,提取标记的痕迹以及识别该痕迹在产品展示页面中的位置,所获取的标记用于识别用户所表达的问题信息,所记录标记的痕迹和位置可以间接描述问题信息中所涉及的文本信息。
例如,标记的痕迹中可以通过文字笔迹、符号笔迹和标记图形展示用户所表述文本信息;另外通过标记的位置与产品展示的页面上内容信息的结合,也可以识别标记所指代的文本信息;或者也可以通过标记的位置分析所识别的文本信息的语义。
步骤S240:将问题信息输入问答模型,获取并在产品展示的页面输出目标答案,其中,问答模型为用于解答页面所展示产品的问题的模型。
本步骤中,将所识别的问题信息输入问答模型,问答模型具有智能的阅读理解功能,问答模型可以解读该页面所展示的内容信息,结合该页面所展示的内容信息对问题信息进行关联和分析,得到用于解答该问题信息的目标答案,即获取目标答案。
本步骤中还可以在该产品展示的页面内输出目标答案,目标页面可以在页面中局部区域进行输出,目标答案与产品展示的内容信息可以在同一页面内进行对比阅读,满足在阅读产品展示的页面的同时,让用户接收该目标答案,便于用户快速阅读和理解目标答案和产品展示的内容信息,提高阅读效率。
上述问答信息的处理方法,可以方便和简易地在当前的产品展示的页面上输入用于提问问题的标记,根据标记和页面的内容信息来识别所提出的问题信息,简便了提出问题信息的过程,可以提高问答过程的效率。而且通过向具有阅读理解功能的智能的问答模型输入问题信息,可以获得与页面的内容信息关系密切的目标答案,提高目标答案的准确度。
进一步地,识别问题信息之前,或者在输出目标答案之前,或者在输出目标答案时,可以在该页面内局部区域中显示已识别的问题信息,便于根据问题信息理解目标答案。
进一步地,目标答案的输出形式可以根据用户选择的或者预设的答案输出模型来决定。在一种实施方式下,步骤S240中在产品展示的页面输出目标答案的步骤,可以包括:
(1)获取预设的答案输出模式。
本步骤中,先查找预设的答案输出模式,确定用户选择的或者预设的答案输出模式对应的信息格式,输出该信息格式下的目标答案。下面将根据答案输出模式对应的信息格式将目标答案进行转换。
(2)在答案输出模式为语音输出时,向页面返回目标答案对应的语音信号。
本步骤中,语音输出的模式下可以通过语音播报该目标答案,用户可以边听边浏览页面的方式理解目标答案,保持停留在当前浏览的产品展示的页面,减少页面的跳转。可以在产品展示的页面设置的某个位置中设置用于触发播放该目标答案对应的语音信号的语音控件。
进一步地,该语音控件还可以用于采集用户输入的语音信号。在获得该输入的语音信号后,通过语音识别获取该输入的语音信号中包含的问题信息,并执行将问题信息输入问答模型,获取并在产品展示的页面输出目标答案的步骤,完成通过语音输入的方式,进行问题提问和问题解答。
(3)在答案输出模式为文本输出时,在页面中显示目标答案对应的文本信息。
本步骤中,文本输出的模式下可以通过文本信息显示该目标答案,在产品展示的页面中局部位置上设置文本显示窗口,该文本显示窗口可以用于显示该目标答案,同时显示产品展示的内容信息,可以在同一个页面内显示目标答案和产品展示的内容信息,减少页面的跳转。
上述问答信息的处理方法,可以提供多种的答案输出模式,按照预设的答案输出模式输出对应信息格式下的目标答案,适应用户的阅读习惯;还可以在输出目标答案时候减少或避免页面的跳转,便于结合产品展示的内容信息理解目标答案。
在一个实施例中,如图3所示,图3为一个实施例中识别问题信息的流程图,步骤S230中根据标记的痕迹和位置以及产品展示的页面识别用户的问题信息的步骤,可以包括:
步骤S231:根据标记的痕迹和位置将标记信息划分为若干个记号。
某次输出的标记内可以包含有若干个不连续拖动或触摸的痕迹,但是这些痕迹之间存在关联关系,例如在局部区域的痕迹组合后可以表现出文字或符号的形状,有的痕迹还可以表现出书面笔记的图形。在标记中可以将一组具有关联关系的痕迹划分为一个记号。
本步骤中,可以通过标记的痕迹和位置判断痕迹之间的关联关系或特 征。不同类型的记号具有特定的特征,以文字的记号为例,可以构成文字的痕迹之间位置会相对紧密,而且痕迹之间常常会有重叠;以下划线为例,痕迹的长度会相对较长;以圆圈标记为例,痕迹的曲度和曲率会较大。然后根据痕迹之间的关联关系将标记中的痕迹划分为若干个记号。
步骤S232:识别各记号的记号类型。
本步骤中,可以根据记号中痕迹的特征或痕迹之间的关联关系,判断记号的记号类型,识别各个记号的记号类型。
步骤S233:根据产品展示的页面和各记号的记号类型,分别一一识别各记号对应的文字信息。
各个记号中会包含所指代的文字信息,各个记号的记号类型进一步识别其所指代的文字。由于记号是在产品展示的页面中产生的,有的记号需要结合页面中的内容信息查找所指代的文字信息,因此还需要结合产品展示的页面进行各记号对应的文字信息的识别。
具体地,根据产品展示的页面和各记号的记号类型,分别一一识别各记号对应的文字信息的步骤,可以包括:
A1:在记号的记号类型为文字轨迹时,将记号转换为笔迹图像,识别笔迹图像中的文字信息,将笔迹图像中的文字信息作为记号对应的文字信息。
本步骤中,当记号属于文字轨迹,记号的痕迹叠加在一起可以直观地观察到该记号相应的文字或符号,再利用图像处理识别该文字或符号,将所识别的文字或符号作为该记号对应的文字信息。
A2:在记号的类型为下划线或圆圈标记时,根据记号在产品展示的页面中的位置,从产品展示的文本内容中提取位置对应的文本信息,将位置对应的文本信息作为记号对应的文字信息。
本步骤中,图形记号可以包括下划线或圆圈,根据图形记号查找显示在该图形记号的位置上的文字信息。
在其中一种实施方式中,可以根据下划线的位置确定文字信息所在的行,根据下划线的首尾位置从该行中确定文字信息的起始位置,在从产品展示的页面中的内容信息中提取该文字信息;另外也可以根据圆圈的位置确定圆圈所覆盖的文字信息,将该被覆盖的文字信息从产品展示的页面中的内容信息中提取。
在其中一种实施方式中,也可以在产品展示的页面中截取图形记号所在位置处的产品展示图像,通过图像处理识别该截取的产品展示图像中的文字和符号,将识别的文字和符号作为记号对应的文字信息。
上述识别各记号对应的文字信息的实施方式,可以有效且准确地将标记转换为文字信息;获得的文字信息可以被计算机和问答模型所识别,以便于后续调用问答模型,提高问题回答的效率和准确度。
下面将以一个具体具有标记的产品展示的页面描述如何识别各记号对应的文字信息,如图4所示,图4为一个实施例中带标记的产品展示的 页面图。其中,记号1、记号2、记号3和记号4作为该产品展示页面的标记。对于记号1、记号2和记号4,分别将记号转换为图像,并进行图像处理,识别出文字“多”、文字“久”和符号“?”。对于记号3,由于记号3的长度较长,记号3偏离水平线的角度较低,可以判定记号3的类型为下划线;获取记号3相对于产品展示页面的位置以及记号3的首尾位置,从页面的内容信息中提取出“对等待期后本主险合同到期日前发生的且延续至本主险合同到期后30日内的”的文字信息。
步骤S234:根据文字信息生成问题信息。
本步骤中,根据自然语言的语法规则对文字信息进行组合和处理,生成符合自然语言语义的问题信息。
具体地,根据文字信息生成问题信息的步骤,可以包括:
B1:识别文字信息的文字和标点符号。
B2:获取文字、标点符号与记号之间的记号位置。
B3:根据文字、标点符号和记号位置,结合语义生成问题信息。
以图4的带标记的产品展示的页面为例进行说明:由于文字信息“多”和“久”的位置解决,可以将文字信息组合为“多久”,同时还能满足自然语言的语义。将“?”作为标点符号,根据自然语言的语法规律,“?”可以方式在问题信息的结束处。“对等待期后本主险合同到期日前发生的且延续至本主险合同到期后30日内的”中包含具有时间含义的词语,可以判断这段文字信息与时间有关,而“多久”也与时间有关,“多久”位于“对等待期后……”的左侧,可以组合得到“对等待期后本主险合同到期日前发生的且延续至本主险合同到期后30日内是多久?”或者“多久对等待期后本主险合同到期日前发生的且延续至本主险合同到期后30日内的”,根据自然语言的语法规则进行处理,对组合后的文本添加可以用于连贯表达意思的词语,计算具有最大可能性的语句,将该最大可能性的语句作为问题信息。
上述生成问题信息的实施例中,可以人工智能地识别用户表达的问题信息,提高问题信息输入的效率和智能化程度。
在一个实施例中,产品展示的页面用于显示保险条款,问答模型用于解答问题信息在保险条款的要求下的答案。
以图4的带标记的产品展示的页面为例进行说明,产品展示的页面显示保险条款,但是保险条款的专业性很强,阅读起来常常出现难以理解的情况,要表示清楚想要提问的问题也是很困难的,例如图4的产品展示的页面中“对等待期后本主险合同到期日前发生的且延续至本主险合同到期后30日内的”中包含多个术语,用户当需要对该条款进行提问,需要复述该条款,往返条款页面和客服对话框,这对于不理解条款的用户来时存在一定描述问题的困难,也大大复杂了用户提问的过程。
本实施例中,当在该条款处进行标记时,就可以自动根据标记识别该条款和生成客户关于该条款的问题,简便了用户提出问题的流程。而且问 答模型可以具有自然语言的阅读理解能力,结合该保险产品的条款的实际语义进行分析后,对问题信息进行解答,所输出的目标答案与该保险条款的关联性强,目标答案的准确性高。
上述问答信息的处理方法,可以提高解答保险条例问题的准确性,提高保险产品的客服质量。
在另一个实施例中,如图5所示,图5为另一个实施例中问答信息的处理方法的流程图。本实施例中提出了一种问答信息的处理方法,具体可以包括以下步骤:
步骤S510:在保险产品的保险条例页面上设置标记输入控件。其中,标记输入控件用于记录并显示在保险条例页面上产生的触摸轨迹和/或鼠标轨迹。
步骤S520:当标记输入控件被触发时,获取用户在保险条例页面上所输入的触摸轨迹和/或鼠标轨迹,并记录触摸轨迹和/或鼠标轨迹的痕迹和位置。
步骤S530:根据触摸轨迹和/或鼠标轨迹的痕迹和位置以及保险条例页面,识别用户的保险产品问题。
具体地,根据触摸轨迹和/或鼠标轨迹的痕迹和位置将其划分为若干个记号。识别各记号的记号类型。
根据保险条例页面的保险条例和各记号的记号类型,分别一一识别各记号对应的文字信息。在记号的记号类型为文字轨迹时,将记号转换为笔迹图像,识别笔迹图像中的文字信息,将笔迹图像中的文字信息作为记号对应的文字信息;在记号的类型为下划线或圆圈标记时,根据记号在产品展示的页面中的位置,从保险条例的文本内容中提取位置对应的文本信息,将位置对应的文本信息作为记号对应的文字信息。
根据文字信息生成保险产品问题。识别文字信息的文字和标点符号;获取文字、标点符号与记号之间的记号位置;根据文字、标点符号和记号位置,结合语义生成保险产品问题。
步骤S540:问答模型用于结合当前浏览的保险条例解答保险产品问题的答案,将保险产品问题输入问答模型,获取并在保险条例页面输出目标答案。
在保险条例页面输出目标答案的步骤,具体可以包括:获取预设的答案输出模式;在答案输出模式为语音输出时,向保险条例页面返回目标答案对应的语音信号;在答案输出模式为文本输出时,在保险条例页面中显示目标答案对应的文本信息。
上述问答信息的处理方法,可以方便和简易地在保险条例页面内输入用于反映保险产品问题的标记,根据标记和保险条例的内容信息来识别所提出的保险产品问题,简便了提出保险产品问题的过程,可以提高保险产品问题的问答过程的效率。而且问答模型具有阅读和理解保险条例功能,向问答模型输入保险产品问题,可以获得与保险条例的内容信息关系密切 的目标答案,提高目标答案的准确度。
在一个实施例中,如图6所示,图6为一个实施例中问答信息的处理系统的结构示意图,本实施例中提出了一种问答信息的处理系统,该问答信息的处理系统具体可以包括设置模块610、获取模块620、识别模块630和输出模块640,其中:
设置模块610,用于在产品展示的页面上设置标记输入控件,其中,标记输入控件用于记录并显示在页面上产生的标记的痕迹。
设置模块610在产品展示的页面上设置有标记输入控制,在浏览当前的产品展示的页面的过程中,可以在当前页面直接触发该标记输入控件,无需跳转至其他页面来查找该标记输入控件。
获取模块620,用于当标记输入控件被触发时,获取用户在页面上所输入的标记,并记录标记的痕迹和位置。
用户在浏览产品展示的页面中触摸移动的痕迹或鼠标移动的痕迹急可以在页面上产生标记的痕迹。获取模块620在控件触发后,可以记录用户在浏览产品展示的页面中触摸移动的痕迹或鼠标移动的痕迹,将这些移动的痕迹作为所输入的标记。
识别模块630,用于根据标记的痕迹和位置以及产品展示的页面识别用户的问题信息。
识别模块630提取标记的痕迹以及识别该痕迹在产品展示页面中的位置,所获取的标记用于识别用户所表达的问题信息,所记录标记的痕迹和位置可以间接描述问题信息中所涉及的文本信息。
输出模块640,用于将问题信息输入问答模型,获取并在产品展示的页面输出目标答案,其中,问答模型为用于解答页面所展示产品的问题的模型。
输出模块640将所识别的问题信息输入问答模型,问答模型具有智能的阅读理解功能,问答模型可以解读该页面所展示的内容信息,结合该页面所展示的内容信息对问题信息进行关联和分析,得到用于解答该问题信息的目标答案,即获取目标答案。
上述问答信息的处理系统,可以方便和简易地在当前的产品展示的页面上输入用于提问问题的标记,根据标记和页面的内容信息来识别所提出的问题信息,简便了提出问题信息的过程,可以提高问答过程的效率。而且通过向具有阅读理解功能的智能的问答模型输入问题信息,可以获得与页面的内容信息关系密切的目标答案,提高目标答案的准确度。
关于问答信息的处理系统的具体限定可以参见上文中对于问答信息的处理方法的限定,在此不再赘述。上述问答信息的处理系统中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中,也可以以软件形式存储于计算机设备中的存储器中,以便于处理器调用执行以上各个模块对应的操作。
如图7所示,图7为一个实施例中计算机设备的内部结构示意图。该计算机设备包括通过系统总线连接的处理器、非易失性存储介质、存储器和网络接口。其中,该计算机设备的非易失性存储介质存储有操作系统、数据库和计算机可读指令,数据库中可存储有控件信息序列,该计算机可读指令被处理器执行时,可使得处理器实现一种问答信息的处理方法。该计算机设备的处理器用于提供计算和控制能力,支撑整个计算机设备的运行。该计算机设备的存储器中可存储有计算机可读指令,该计算机可读指令被处理器执行时,可使得处理器执行一种问答信息的处理方法。该计算机设备的网络接口用于与终端连接通信。本领域技术人员可以理解,图7中示出的结构,仅仅是与本申请方案相关的部分结构的框图,并不构成对本申请方案所应用于其上的计算机设备的限定,具体的计算机设备可以包括比图中所示更多或更少的部件,或者组合某些部件,或者具有不同的部件布置。
在一个实施例中,提出了一种计算机设备,计算机设备包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序,处理器执行计算机程序时实现上述任一实施例的问答信息的处理方法的步骤。
在一个实施例中,提出了一种存储有计算机可读指令的存储介质,该存储介质可以是非易失性的,也可以是易失性的,该计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行上述任一实施例的问答信息的处理方法的步骤。
应该理解的是,虽然附图的流程图中的各个步骤按照箭头的指示依次显示,但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明,这些步骤的执行并没有严格的顺序限制,其可以以其他的顺序执行。而且,附图的流程图中的至少一部分步骤可以包括多个子步骤或者多个阶段,这些子步骤或者阶段并不必然是在同一时刻执行完成,而是可以在不同的时刻执行,其执行顺序也不必然是依次进行,而是可以与其他步骤或者其他步骤的子步骤或者阶段的至少一部分轮流或者交替地执行。
以上所述仅是本申请的部分实施方式,应当指出,对于本技术领域的普通技术人员来说,在不脱离本申请原理的前提下,还可以做出若干改进和润饰,这些改进和润饰也应视为本申请的保护范围。

Claims (20)

  1. 一种问答信息的处理方法,其中,包括如下步骤:
    在产品展示的页面上设置标记输入控件,其中,所述标记输入控件用于记录并显示在所述页面上产生的标记的痕迹;
    当所述标记输入控件被触发时,获取用户在所述页面上所输入的标记,并记录所述标记的痕迹和位置;
    根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息;
    将所述问题信息输入问答模型,获取并在所述产品展示的页面输出目标答案,其中,所述问答模型为用于解答所述页面所展示产品的问题的模型。
  2. 根据权利要求1所述的问答信息的处理方法,其中,所述标记包括触摸轨迹或鼠标轨迹。
  3. 根据权利要求1所述的问答信息的处理方法,其中,所述根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息的步骤,包括:
    根据所述标记的痕迹和位置将所述标记信息划分为若干个记号;
    识别各所述记号的记号类型;
    根据所述产品展示的页面和各所述记号的记号类型,分别一一识别各所述记号对应的文字信息;
    根据所述文字信息生成所述问题信息。
  4. 根据权利要求3所述的问答信息的处理方法,其中,所述根据所述产品展示的页面和各所述记号的记号类型,分别一一识别各所述记号对应的文字信息的步骤,包括:
    在所述记号的记号类型为文字轨迹时,将所述记号转换为笔迹图像,识别所述笔迹图像中的文字信息,将所述笔迹图像中的文字信息作为所述记号对应的文字信息;
    在所述记号的类型为下划线或圆圈标记时,根据所述记号在所述产品展示的页面中的位置,从所述产品展示的文本内容中提取所述位置对应的文本信息,将所述位置对应的文本信息作为所述记号对应的文字信息。
  5. 根据权利要求3所述的问答信息的处理方法,其中,所述根据所述文字信息生成所述问题信息的步骤,包括:
    识别所述文字信息的文字和标点符号;
    获取所述文字、所述标点符号与所述记号之间的记号位置;
    根据所述文字、所述标点符号和所述记号位置,结合语义生成所述问题信息。
  6. 根据权利要求1所述的问答信息的处理方法,其中,所述在所述产品展示的页面输出目标答案的步骤,包括:
    获取预设的答案输出模式;
    在所述答案输出模式为语音输出时,向所述页面返回所述目标答案对应的语音信号;
    在所述答案输出模式为文本输出时,在所述页面中显示所述目标答案对应的文本信息。
  7. 根据权利要求1所述的问答信息的处理方法,其中,所述产品展示的页面用于显示保险条款,所述问答模型用于解答所述问题信息在所述保险条款的要求下的答案。
  8. 一种问答信息的处理系统,其中,包括:
    设置模块,用于在产品展示的页面上设置标记输入控件,其中,所述标记输入控件用于记录并显示在所述页面上产生的标记的痕迹;
    获取模块,用于当所述标记输入控件被触发时,获取用户在所述页面上所输入的标记,并记录所述标记的痕迹和位置;
    识别模块,用于根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息;
    输出模块,用于将所述问题信息输入问答模型,获取并在所述产品展示的页面输出目标答案,其中,所述问答模型为用于解答所述页面所展示产品的问题的模型。
  9. 一种计算机设备,包括存储器和处理器,所述存储器中存储有计算机可读指令,其中,所述计算机可读指令被所述处理器执行时,使得所述处理器执行如下步骤:
    在产品展示的页面上设置标记输入控件,其中,所述标记输入控件用于记录并显示在所述页面上产生的标记的痕迹;
    当所述标记输入控件被触发时,获取用户在所述页面上所输入的标记,并记录所述标记的痕迹和位置;
    根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息;
    将所述问题信息输入问答模型,获取并在所述产品展示的页面输出目标答案,其中,所述问答模型为用于解答所述页面所展示产品的问题的模型。
  10. 如权利要求9所述的计算机设备,其中,所述标记包括触摸轨迹或鼠标轨迹。
  11. 如权利要求9所述的计算机设备,其中,所述计算机可读指令被所述处理器执行时,使得所述处理器执行所述根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息的步骤,包括:
    根据所述标记的痕迹和位置将所述标记信息划分为若干个记号;
    识别各所述记号的记号类型;
    根据所述产品展示的页面和各所述记号的记号类型,分别一一识别各所述记号对应的文字信息;
    根据所述文字信息生成所述问题信息。
  12. 如权利要求11所述的计算机设备,其中,所述计算机可读指令被所述处理器执行时,使得所述处理器执行所述根据所述产品展示的页面和各所述记号的记号类型,分别一一识别各所述记号对应的文字信息的步骤,包括:
    在所述记号的记号类型为文字轨迹时,将所述记号转换为笔迹图像,识别所述笔迹图像中的文字信息,将所述笔迹图像中的文字信息作为所述记号对应的文字信息;
    在所述记号的类型为下划线或圆圈标记时,根据所述记号在所述产品展示的页面中的位置,从所述产品展示的文本内容中提取所述位置对应的文本信息,将所述位置对应的文本信息作为所述记号对应的文字信息。
  13. 如权利要求11所述的计算机设备,其中,所述计算机可读指令被所述处理器执行时,使得所述处理器执行所述根据所述文字信息生成所述问题信息的步骤,包括:
    识别所述文字信息的文字和标点符号;
    获取所述文字、所述标点符号与所述记号之间的记号位置;
    根据所述文字、所述标点符号和所述记号位置,结合语义生成所述问题信息。
  14. 如权利要求9所述的计算机设备,其中,所述计算机可读指令被所述处理器执行时,使得所述处理器执行所述在所述产品展示的页面输出目标答案的步骤,包括:
    获取预设的答案输出模式;
    在所述答案输出模式为语音输出时,向所述页面返回所述目标答案对应的语音信号;
    在所述答案输出模式为文本输出时,在所述页面中显示所述目标答案对应的文本信息。
  15. 一种存储有计算机可读指令的存储介质,其中,所述计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行如下步骤:
    在产品展示的页面上设置标记输入控件,其中,所述标记输入控件用于记录并显示在所述页面上产生的标记的痕迹;
    当所述标记输入控件被触发时,获取用户在所述页面上所输入的标记,并记录所述标记的痕迹和位置;
    根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息;
    将所述问题信息输入问答模型,获取并在所述产品展示的页面输出目标答案,其中,所述问答模型为用于解答所述页面所展示产品的问题的模型。
  16. 如权利要求15所述的存储介质,其中,所述标记包括触摸轨迹或鼠标轨迹。
  17. 如权利要求15所述的存储介质,其中,所述计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行所述根据所述标记的痕迹和位置以及所述产品展示的页面识别所述用户的问题信息的步骤,包括:
    根据所述标记的痕迹和位置将所述标记信息划分为若干个记号;
    识别各所述记号的记号类型;
    根据所述产品展示的页面和各所述记号的记号类型,分别一一识别各所述记号对应的文字信息;
    根据所述文字信息生成所述问题信息。
  18. 如权利要求17所述的存储介质,其中,所述计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行所述根据所述产品展示的页面和各所述记号的记号类型,分别一一识别各所述记号对应的文字信息的步骤,包括:
    在所述记号的记号类型为文字轨迹时,将所述记号转换为笔迹图像,识别所述笔迹图像中的文字信息,将所述笔迹图像中的文字信息作为所述记号对应的文字信息;
    在所述记号的类型为下划线或圆圈标记时,根据所述记号在所述产品展示的页面中的位置,从所述产品展示的文本内容中提取所述位置对应的文本信息,将所述位置对应的文本信息作为所述记号对应的文字信息。
  19. 如权利要求17所述的存储介质,其中,所述计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行所述根据所述文字信息生成所述问题信息的步骤,包括:
    识别所述文字信息的文字和标点符号;
    获取所述文字、所述标点符号与所述记号之间的记号位置;
    根据所述文字、所述标点符号和所述记号位置,结合语义生成所述问题信息。
  20. 如权利要求15所述的存储介质,其中,所述计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行所述在所述产品展示的页面输出目标答案的步骤,包括:
    获取预设的答案输出模式;
    在所述答案输出模式为语音输出时,向所述页面返回所述目标答案对应的语音信号;
    在所述答案输出模式为文本输出时,在所述页面中显示所述目标答案对应的文本信息。
PCT/CN2020/086128 2019-05-24 2020-04-22 问答信息的处理方法、系统、计算机设备和存储介质 Ceased WO2020238498A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910440492.9A CN110308947A (zh) 2019-05-24 2019-05-24 问答信息的处理方法、系统、计算机设备和存储介质
CN201910440492.9 2019-05-24

Publications (1)

Publication Number Publication Date
WO2020238498A1 true WO2020238498A1 (zh) 2020-12-03

Family

ID=68074842

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/086128 Ceased WO2020238498A1 (zh) 2019-05-24 2020-04-22 问答信息的处理方法、系统、计算机设备和存储介质

Country Status (2)

Country Link
CN (1) CN110308947A (zh)
WO (1) WO2020238498A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110308947A (zh) * 2019-05-24 2019-10-08 深圳壹账通智能科技有限公司 问答信息的处理方法、系统、计算机设备和存储介质
CN111951133B (zh) * 2020-07-31 2024-03-22 北京伴我学教育科技有限公司 辅助解题的方法、装置以及存储介质
CN114238787B (zh) * 2020-08-31 2024-03-29 腾讯科技(深圳)有限公司 答题处理方法、装置
CN113849690A (zh) * 2021-09-27 2021-12-28 联想(北京)有限公司 问答处理方法、装置及计算机设备
CN118354113B (zh) * 2024-04-09 2025-08-05 北京达佳互联信息技术有限公司 讲解信息的展示方法、装置、设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103885966A (zh) * 2012-12-20 2014-06-25 北京京东尚科信息技术有限公司 电子商务交易平台中的问答交互方法和系统
US20170068854A1 (en) * 2015-09-03 2017-03-09 Microsoft Technology Licensing, Llc Interacting with an Assistant Component Based on Captured Stroke Information
CN108345692A (zh) * 2018-03-16 2018-07-31 北京京东尚科信息技术有限公司 一种自动问答方法和系统
CN108763561A (zh) * 2018-06-04 2018-11-06 深圳市零度智控科技有限公司 电子书推送方法、装置及计算机可读存储介质
CN110308947A (zh) * 2019-05-24 2019-10-08 深圳壹账通智能科技有限公司 问答信息的处理方法、系统、计算机设备和存储介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104331441A (zh) * 2014-10-24 2015-02-04 北京奇虎科技有限公司 基于搜索引擎提供问题答案的方法和装置
CN105139698B (zh) * 2015-07-31 2018-03-27 广东小天才科技有限公司 一种点读机的信息录入方法及装置
CN109285030A (zh) * 2018-08-29 2019-01-29 深圳壹账通智能科技有限公司 产品推荐方法、装置、终端及计算机可读存储介质
CN109697228A (zh) * 2018-12-13 2019-04-30 平安科技(深圳)有限公司 智能问答方法、装置、计算机设备及存储介质
CN109670088B (zh) * 2018-12-13 2024-12-17 平安科技(深圳)有限公司 语音问答交互方法、装置、计算机设备及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103885966A (zh) * 2012-12-20 2014-06-25 北京京东尚科信息技术有限公司 电子商务交易平台中的问答交互方法和系统
US20170068854A1 (en) * 2015-09-03 2017-03-09 Microsoft Technology Licensing, Llc Interacting with an Assistant Component Based on Captured Stroke Information
CN108345692A (zh) * 2018-03-16 2018-07-31 北京京东尚科信息技术有限公司 一种自动问答方法和系统
CN108763561A (zh) * 2018-06-04 2018-11-06 深圳市零度智控科技有限公司 电子书推送方法、装置及计算机可读存储介质
CN110308947A (zh) * 2019-05-24 2019-10-08 深圳壹账通智能科技有限公司 问答信息的处理方法、系统、计算机设备和存储介质

Also Published As

Publication number Publication date
CN110308947A (zh) 2019-10-08

Similar Documents

Publication Publication Date Title
WO2020238498A1 (zh) 问答信息的处理方法、系统、计算机设备和存储介质
US10733197B2 (en) Method and apparatus for providing information based on artificial intelligence
CN109348275B (zh) 视频处理方法和装置
US10192544B2 (en) Method and system for constructing a language model
US20200134398A1 (en) Determining intent from multimodal content embedded in a common geometric space
CN111459977B (zh) 自然语言查询的转换
WO2021139247A1 (zh) 医学领域知识图谱的构建方法、装置、设备及存储介质
WO2019153685A1 (zh) 文本处理方法、装置、计算机设备和存储介质
CN110956026B (zh) 法律文书生成方法、生成装置和电子设备
US20170351371A1 (en) Touch interaction based search method and apparatus
WO2021068932A1 (zh) 基于电子书的实体关联信息的展示方法及电子设备
CN110598217B (zh) 一种点读内容的识别方法、装置、家教机及存储介质
CN109343696B (zh) 一种电子书的评论方法、装置及计算机可读存储介质
WO2020199600A1 (zh) 情感极性分析方法及相关装置
WO2019075819A1 (zh) 一种判卷方法及判卷系统
US20210141865A1 (en) Machine learning based tenant-specific chatbots for performing actions in a multi-tenant system
CN111581972A (zh) 文本中症状和部位对应关系识别方法、装置、设备及介质
CN101561725B (zh) 一种手写快速输入的方法和系统
CN114925206A (zh) 人工智能体、语音信息识别方法、存储介质和程序产品
US20250173917A1 (en) Video generation method and apparatus, computer device, and storage medium
CN114637831A (zh) 基于语义分析的数据查询方法及其相关设备
EP3292480A1 (en) Techniques to automatically generate bookmarks for media files
WO2021254046A1 (zh) 一种信息查询方法及装置
US20210056149A1 (en) Search system, search method, and program
CN117579888A (zh) 视频字幕提取方法、装置、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: 20812825

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: 20812825

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 21/03/2022)

122 Ep: pct application non-entry in european phase

Ref document number: 20812825

Country of ref document: EP

Kind code of ref document: A1