Search

CN-122001840-A - Prompt generation method, device, equipment, storage medium and program product

CN122001840ACN 122001840 ACN122001840 ACN 122001840ACN-122001840-A

Abstract

The present application relates to a hint generation method, apparatus, computer device, computer readable storage medium and computer program product. The method comprises the steps of generating a first session state based on first interaction information obtained through multiple rounds of interaction, obtaining second interaction information in different data forms after a first interaction prompt is displayed based on the first session state, enabling the different data forms to comprise data forms input by different devices, analyzing and converting formats of the second interaction information to obtain state increment, conducting consistency processing on the state increment and the state information in the first session state to obtain a second session state, and displaying the second interaction prompt based on the second session state. By adopting the method, the irreversible distortion of the conversation state caused by the second interactive information input by the multisource asynchronization can be avoided, so that the consistency of the conversation state is ensured, the consistency of the interactive prompt can be ensured, and the inconsistency of the interactive prompt before and after the state evolution is avoided.

Inventors

  • HUANG HAITAO

Assignees

  • 深圳智界探索科技有限公司

Dates

Publication Date
20260508
Application Date
20260211

Claims (10)

  1. 1. A hint generation method, the method comprising: Generating a first session state based on first interaction information obtained by the multi-round interaction; After a first interaction prompt is displayed based on the first session state, second interaction information of different data forms is acquired, wherein the different data forms comprise data forms input by different equipment; analyzing and converting the second interaction information to obtain a state increment; Consistency processing is carried out on the state increment and the state information in the first session state, so that a second session state is obtained; and displaying a second interaction prompt based on the second session state.
  2. 2. The method of claim 1, wherein the parsing and format converting the second interaction information to obtain a state increment includes: Converting the second interaction information into each interaction event containing interaction content and metadata, wherein the metadata represents the acquisition mode and time of the second interaction information; determining a status value of the content type based on the interactive content of the same content type; Associating each state value with the metadata to obtain each state increment; And performing consistency processing on the state increment and the state information in the first session state to obtain a second session state, wherein the method comprises the following steps: Selecting an available state value based on the time of each interaction event in the state increment of different acquisition modes; and adjusting the state information in the first session state according to the state increment containing the available state value to obtain a second session state.
  3. 3. The method of claim 2, wherein the determining a status value for the content type based on the interactive content of the same content type comprises: splitting the composite value into a plurality of state values of the content types in the case that the interactive content is determined to comprise the composite value; in the case that the interactive content of the same content type covers a plurality of moments, determining a state value of the same content type at each moment.
  4. 4. The method of claim 1, wherein the state delta comprises metadata associated with a state value of the interaction event and the state value; And performing consistency processing on the state increment and the state information in the first session state to obtain a second session state, wherein the method comprises the following steps: Under the condition that the state value belongs to a non-conflict content type based on the metadata, according to the content type to which the state value belongs, the state value is used for covering or merging the state information in the first session state, so that the second session state is obtained; And when the state value is determined to belong to a conflict content type based on each metadata, obtaining the second session state based on the state value of the same content type and the state information in the first session state respectively.
  5. 5. The method of claim 4, wherein the deriving the second session state based on the state value and state information in the first session state, respectively, of the same content type comprises: determining different candidate values for the same content type based on the state value and state information in the first session state for the same content type; Determining a cause of the difference of the different candidate values; the displaying a second interaction prompt based on the second session state includes: And generating and displaying a second interaction prompt according to the difference reason.
  6. 6. The method of claim 1, wherein generating the first session state based on the first interaction information resulting from the multiple rounds of interactions comprises: binding the interaction process identification of each round of interaction with input information to obtain first interaction information; Binding the first interaction information with at least one object identifier to obtain a first session state of at least one object.
  7. 7. A hint generation apparatus, the apparatus comprising: The session analysis module is used for generating a first session state based on first interaction information obtained by multiple rounds of interaction; The interactive input module is used for acquiring second interactive information with different data forms after the first interactive prompt is displayed based on the first session state, wherein the different data forms comprise data forms input by different equipment; The session analysis module is used for analyzing and converting the format of the second interaction information to obtain a state increment; The session analysis module is used for carrying out consistency processing on the state increment and the state information in the first session state to obtain a second session state; and the prompt output module is used for displaying a second interaction prompt based on the second session state.
  8. 8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
  9. 9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
  10. 10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.

Description

Prompt generation method, device, equipment, storage medium and program product Technical Field The present application relates to the field of artificial intelligence technology, and in particular, to a method, an apparatus, a computer device, a computer readable storage medium, and a computer program product for generating a hint. Background With the continuous progress of technology, the application scene of artificial intelligence is becoming more complex. In many practical applications, the current intelligent question-answering system has relatively strong context awareness capability and can feed back through context interaction information. However, the context sensing range is limited, and the context processing is only performed based on the user input of each round of interaction process, so that the terminal needs to repeatedly input related information for a plurality of times in the multi-round interaction process, thereby easily causing that generated interaction prompts lack of consistency between the previous round and the next round, and the interaction consistency for generating the interaction prompts is poor, and the interaction prompts may be inconsistent in state evolution. Disclosure of Invention In view of the foregoing, it is desirable to provide a method, an apparatus, a computer device, a computer readable storage medium, and a computer program product for generating a hint, which can ensure consistency of an interactive hint, and avoid inconsistency of the interactive hint in state evolution. In a first aspect, the present application provides a method for generating a hint, including: Generating a first session state based on first interaction information obtained by the multi-round interaction; After a first interaction prompt is displayed based on the first session state, second interaction information of different data forms is acquired, wherein the different data forms comprise data forms input by different equipment; analyzing and converting the second interaction information to obtain a state increment; Consistency processing is carried out on the state increment and the state information in the first session state, so that a second session state is obtained; and displaying a second interaction prompt based on the second session state. In one embodiment, the parsing and format converting the second interaction information to obtain a state increment includes: Converting the second interaction information into each interaction event containing interaction content and metadata, wherein the metadata represents the acquisition mode and time of the second interaction information; determining a status value of the content type based on the interactive content of the same content type; Associating each state value with the metadata to obtain each state increment; And performing consistency processing on the state increment and the state information in the first session state to obtain a second session state, wherein the method comprises the following steps: Selecting an available state value based on the time of each interaction event in the state increment of different acquisition modes; and adjusting the state information in the first session state according to the state increment containing the available state value to obtain a second session state. In one embodiment, the determining the status value of the content type based on the interactive content of the same content type includes: splitting the composite value into a plurality of state values of the content types in the case that the interactive content is determined to comprise the composite value; In the case that the interactive content of the same content type covers a plurality of moments, determining a state value of the same content type at each moment. In one embodiment, the state delta includes metadata associated with a state value of the interaction event and the state value; And performing consistency processing on the state increment and the state information in the first session state to obtain a second session state, wherein the method comprises the following steps: Under the condition that the state value belongs to a non-conflict content type based on the metadata, according to the content type to which the state value belongs, the state value is used for covering or merging the state information in the first session state, so that the second session state is obtained; And when the state value is determined to belong to the conflict content type based on each metadata, obtaining the second session state based on the state value and the state information in the first session state respectively. In one embodiment, the obtaining the second session state based on the state value and the state information in the first session state of the same content type respectively includes: Determining different candidate values of the same content type based on the state value of the same content type and state informa