Method for relieving robot session failure by utilizing interpretive interaction

Abstract

The invention belongs to the technical field of natural language processing, and particularly relates to a method for relieving a robot session failure by utilizing explanatory interaction, which comprises the following steps of: monitoring a session process between a robot and a user in real time by utilizing a natural language processing technology and an emotion analysis algorithm; generating corresponding explanatory interaction content; personalized explanatory interaction content is provided for each user by analyzing the interaction history and preference setting information of the user; and feedback suggestions of the user on the explanatory interaction are collected, system optimization is carried out according to feedback information, and the quality and effect of the explanatory interaction are continuously improved. According to the method, an explanatory interaction mechanism is introduced, so that the robot can provide detailed explanation and description for the user when the session failure occurs, and the user is helped to better understand the behavior and state of the robot. Meanwhile, by continuously optimizing the content and form of the explanatory interaction, the satisfaction degree and credibility of the user can be improved, and effective man-machine communication is promoted.

Description

Technical Field

[0001] The present invention belongs to the technical field of natural language processing, and particularly relates to a method for alleviating robot conversation failure by means of interpretive interaction. Background Art

[0002] In recent years, with the rapid development of artificial intelligence technology, robots have been able to conduct natural language interactions with users in many scenarios and complete various tasks. However, the situation of robot conversation failure still occurs from time to time, which is mainly due to the limited processing capabilities of robots in aspects such as complex contexts, polysemous words, and emotional understanding.

[0003] Conversation failure not only affects the user experience but also may lead to a decrease in users' trust in robot technology. Therefore, how to effectively alleviate robot conversation failure and improve the conversation interaction ability of robots has become an urgent problem to be solved in the current field of robot technology.

[0004] Currently, there are mainly two methods for dealing with robot conversation failure: one is to optimize the natural language processing technology of the robot to improve its ability to understand and parse user inputs; the other is to design a more reasonable interaction process to reduce the possibility of conversation failure. However, these methods can only alleviate the problem of conversation failure to a certain extent and cannot fundamentally solve users' confusion and dissatisfaction with robot behavior. Summary of the Invention

[0005] In view of the above technical problem that the existing methods for dealing with robot conversation failure cannot fundamentally solve users' confusion and dissatisfaction with robot behavior, the present invention provides a method for alleviating robot conversation failure by means of interpretive interaction.

[0006] To solve the above technical problem, the technical solution adopted by the present invention is as follows:

[0007] A method for alleviating robot conversation failure by means of interpretive interaction, comprising the following steps:

[0008] S1. Using natural language processing technology and sentiment analysis algorithms, monitor the conversation process between the robot and the user in real time; when it is detected that there are ambiguous, misunderstood, or unresponsive situations in the conversation content, determine it as a conversation failure and trigger the interpretive interaction generation process;

[0009] S2. Generate corresponding interpretive interaction content according to the specific reasons for the conversation failure, including language understanding errors and task execution difficulties; the interpretive content includes various forms such as text descriptions, voice explanations, images, or video displays to meet the needs of different users.

[0010] S3. Provide personalized explanatory interaction content for each user by analyzing the user's interaction history and preference settings information; consider the user's language habits and cultural background factors to ensure that the explanatory content is easy to understand and accept;

[0011] S4. Collect feedback from users on the explanatory interaction, including satisfaction evaluations and improvement suggestions; optimize the system based on the feedback information to continuously improve the quality and effect of the explanatory interaction.

[0012] The method for real-time monitoring of the conversation process between the robot and the user in S1 is as follows:

[0013] S1.1. Encode the user input text U and the robot response text R using a pre-trained language model to obtain semantic vectors u and r, and calculate the semantic similarity S(u, r);

[0014] S1.2. When S(u, r) < θs, it is determined that the semantics are inconsistent, where θs is a preset threshold;

[0015] Calculate the sentiment score E(U) of the user input using a sentiment analysis model; when E(U) < θe, it is determined that the sentiment is negative, where θe is the sentiment threshold;

[0016] When the robot cannot generate a valid response and the confidence level is lower than the threshold θc, it is determined that the robot cannot respond;

[0017] S1.3. Design a decision formula:

[0018]

[0019] Among them, F(U, R) = 1 indicates that the conversation fails.

[0020] The method for calculating the semantic similarity S(u, r) in S1.1 is as follows:

[0021]

[0022] Among them, S(u, r) represents the semantic consistency between the user input and the robot response.

[0023] The method for generating corresponding explanatory interaction content according to the specific reasons for the conversation failure in S2 is as follows:

[0024] S2.1. Use natural language processing (NLP) technology to detect ambiguities or errors in the user input;

[0025]

[0026] Among them: P(error∣input) represents the probability of error under a given input, P(input∣error) represents the probability of the input in the case of error, P(error) represents the prior probability of error, and P(input) represents the prior probability of the input;

[0027] S2.2. Generate corresponding explanatory interaction content according to the detected error type;

[0028] E = f(error_type, context)

[0029] Among them: E represents the generated explanatory content, error_type represents the error type, and context represents the context information.

[0030] The method for providing personalized explanatory interaction content for each user by analyzing the user's interaction history and preference setting information in S3 is as follows:

[0031] Select one or more of text, voice, image, or video to output the explanatory interaction content according to user preferences;

[0032]

[0033] Among them: Output represents the finally output explanatory content, w i represents the preference weight of the user for the i-th form, and E i represents the explanatory content of the i-th form.

[0034] The method for optimizing the system according to the feedback information in S4 to continuously improve the quality and effect of explanatory interaction is as follows:

[0035] S4.1. Adjust the robot conversation model according to user feedback and failure scenarios, and optimize the explanatory interaction strategy;

[0036] S4.2. Design new explanatory interaction strategies for high-frequency failure scenarios and conduct A / B tests; redeploy the optimized system, continue to collect user feedback, and form a closed-loop optimization process;

[0037] S4.3. Compare the user satisfaction and session failure rate before and after optimization, and evaluate the optimization effect;

[0038] S4.4. Establish a continuous monitoring mechanism, regularly evaluate the system performance, and ensure the continuous improvement of the quality and effect of explanatory interaction.

[0039] The method for optimizing the explanatory interaction strategy in S4.1 is as follows:

[0040] Adjust the model parameter weights according to user satisfaction

[0041] ω new = ω old + α × (Satisfaction score - Mean satisfaction)

[0042] Where α is the learning rate, ω new represents the optimized interpretive interaction strategy parameter, and ω old represents the interpretive interaction strategy parameter before optimization.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0044] By introducing an interpretive interaction mechanism, the present invention enables the robot to provide detailed explanations and descriptions to the user when a conversation fails, helping the user better understand the behavior and state of the robot. At the same time, by continuously optimizing the content and form of the interpretive interaction, the present invention can also improve the user's satisfaction and trust, and promote effective communication between humans and machines. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and those of ordinary skill in the art can also obtain other implementation drawings according to the provided drawings without creative efforts.

[0046] The structures, proportions, sizes, etc. illustrated in this specification are only used to cooperate with the content disclosed in the specification for those skilled in this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention. Therefore, they do not have technical substance significance. Any modification of the structure, change of the proportional relationship, or adjustment of the size should still fall within the scope covered by the technical content disclosed in the present invention without affecting the effects that the present invention can produce and the purposes that can be achieved.

[0047] Figure 1 is the overall architecture diagram of the present invention;

[0048] Figure 2 is the module relationship diagram of the present invention;

[0049] Figure 3 is the flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. These descriptions are only for further explaining the features and advantages of the present invention, rather than limiting the claims of the present invention; based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present application.

[0051] The following will further describe in detail the specific implementation manners of the present invention in conjunction with the accompanying drawings and embodiments. The following embodiments are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0052] A method for alleviating robot conversation failure by using interpretive interaction, as Figures 1-3 shown, includes the following steps:

[0053] Step 1, Conversation failure detection module: Using natural language processing technology and sentiment analysis algorithms, it monitors the conversation process between the robot and the user in real time; when it detects that there are ambiguities, misunderstandings, or unanswerable situations in the conversation content, it determines that the conversation has failed and triggers the interpretive interaction generation process.

[0054] Step 1.1, Encode the user input text U and the robot response text R using a pre-trained language model to obtain semantic vectors u and r, and calculate the semantic similarity S(u, r):

[0055]

[0056] where S(u, r) represents the semantic consistency between the user input and the robot response.

[0057] Step 1.2, When S(u, r) < θs, it is determined that the semantics are inconsistent, where θs is a preset threshold.

[0058] Calculate the sentiment score E(U) of the user input using a sentiment analysis model; when E(U) < θe, it is determined that the sentiment is negative, where θe is a sentiment threshold.

[0059] When the robot cannot generate a valid response and the confidence level is lower than the threshold θc, it is determined that it cannot answer;

[0060] Step 1.3, Design a determination formula:

[0061]

[0062] where F(U, R) = 1 indicates that the conversation has failed.

[0063] Step 2, Explanatory Interaction Generation Module: Generate corresponding explanatory interaction content according to the specific reasons for session failure, including language understanding errors and task execution difficulties; the explanatory content includes various forms such as text descriptions, voice explanations, image or video displays to meet the needs of different users.

[0064] Step 2.1, Use natural language processing NLP technology to detect ambiguities or errors in the user input;

[0065]

[0066] Where: P(error∣input) represents the probability of error given the input, P(input∣error) represents the probability of the input in case of error, P(error) represents the prior probability of error, and P(input) represents the prior probability of the input.

[0067] Step 2.2, Generate corresponding explanatory interaction content according to the detected error type;

[0068] E = f(error_type, context)

[0069] Where: E represents the generated explanatory content, error_type represents the error type, and context represents the context information.

[0070] Step 3, Personalized Explanation Provision Module: Provide personalized explanatory interaction content for each user by analyzing the user's interaction history and preference setting information; consider the user's language habits and cultural background factors to ensure that the explanatory content is easy to understand and accept.

[0071] Select one or more of text, voice, image, or video to output the explanatory interaction content according to the user's preference;

[0072]

[0073] Where: Output represents the finally output explanatory content, w i represents the user's preference weight for the i-th form, and E i represents the explanatory content of the i-th form.

[0074] Step 4, Feedback Collection and Optimization Module: Collect the user's feedback on the explanatory interaction, including satisfaction evaluations and improvement suggestions; optimize the system according to the feedback information to continuously improve the quality and effect of the explanatory interaction.

[0075] Step 4.1, Adjust the robot session model according to the user feedback and failure scenarios, and optimize the explanatory interaction strategy.

[0076] Adjust the model parameter weights according to user satisfaction

[0077] ω new = ω old + α × (satisfaction score - average satisfaction)

[0078] where α is the learning rate, ω new represents the optimized explanatory interaction strategy parameters, and ω old represents the explanatory interaction strategy parameters before optimization.

[0079] Step 4.2: Design a new explanatory interaction strategy for high-frequency failure scenarios and conduct A / B testing; redeploy the optimized system, continue to collect user feedback, and form a closed-loop optimization process.

[0080] Step 4.3: Compare the user satisfaction and session failure rate before and after optimization to evaluate the optimization effect.

[0081] Step 4.4: Establish a continuous monitoring mechanism to regularly evaluate the system performance and ensure the continuous improvement of the quality and effect of explanatory interaction.

[0082] Embodiment

[0083] Taking a domestic service robot as an example, assume that the user issues an instruction to the robot: "Please help me open the window in the bedroom." Due to the window being stuck or motor failure, the robot cannot execute this instruction. At this time, the robot will activate the session failure detection module to identify the current session failure situation.

[0084] Then, the explanatory interaction generation module will generate explanatory content based on the failure reason. For example, the robot can explain to the user via voice: "Sorry, I can't open the window in the bedroom because the window seems to be stuck. I'm trying to solve this problem, please wait a moment." At the same time, the robot can also display a picture of the stuck window on the display screen to help the user better understand the problem.

[0085] If the user expressed a preference for text explanations in previous interactions, the personalized explanation providing module will adjust the explanation form according to this information and explain the reason for the session failure to the user in text form.

[0086] Finally, the feedback collection and optimization module will collect the user's feedback on this explanatory interaction. If the user expresses satisfaction, the system will continue to maintain the current strategy; if the user puts forward improvement suggestions, such as hoping for more detailed explanations or adding other forms of interaction, the system will make corresponding optimizations according to these suggestions.

[0087] Through the above embodiments, the present invention can provide timely, accurate, and personalized explanatory interactions when the robot session fails, effectively alleviating users' confusion and dissatisfaction, and improving the efficiency and satisfaction of human-computer interaction.

[0088] The above only elaborates in detail on the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those of ordinary skill in the art, various changes can be made without departing from the gist of the present invention, and all such changes should be included within the protection scope of the present invention.

Claims

1. A method for alleviating robot conversation failure by using interpretive interaction, characterized in that, It includes the following steps: S1. Use natural language processing technology and sentiment analysis algorithms to monitor the conversation process between the robot and the user in real time; when it is detected that there are ambiguities, misunderstandings or situations where the robot cannot respond in the conversation content, it is determined that the conversation fails, and the explanatory interaction generation process is triggered; S2. Generate corresponding explanatory interaction content according to the specific reasons for the conversation failure, including language understanding errors and task execution difficulties; the explanatory content includes various forms such as text descriptions, voice explanations, images or video displays to meet the needs of different users; S3. Provide personalized explanatory interaction content for each user by analyzing the user's interaction history and preference setting information; Consider the user's language habits and cultural background factors to ensure that the explanatory content is easy to understand and accept; S4. Collect the feedback from users on the explanatory interaction, including satisfaction evaluations and improvement suggestions; Optimize the system according to the feedback information to continuously improve the quality and effect of the explanatory interaction.

2. The method for alleviating robot conversation failure by using interpretive interaction according to claim 1, characterized in that, The method for monitoring the conversation process between the robot and the user in real time in S1 is as follows: S1.

1. Use a pre-trained language model to encode the user input text U and the robot response text R to obtain semantic vectors u and r, and calculate the semantic similarity S(u, r); S1.

2. When S(u, r) < θs, it is determined that the semantics are inconsistent, where θs is a preset threshold; Use a sentiment analysis model to calculate the sentiment score E(U) of the user input; when E(U) < θe, it is determined that the sentiment is negative, where θe is the sentiment threshold; When the robot cannot generate an effective response and the confidence level is lower than the threshold θc, it is determined that the robot cannot respond; S1.

3. Design a determination formula: where F(U, R) = 1 indicates that the conversation fails.

3. The method for alleviating robot conversation failure by using interpretive interaction according to claim 2, wherein, The method for calculating the semantic similarity S(u, r) in S1.1 is as follows: where S(u, r) represents the semantic consistency between the user input and the robot response.

4. A method for alleviating robot conversation failure by using interpretive interaction according to claim 1, characterized in that, The method for generating corresponding explanatory interaction content according to the specific reasons for the conversation failure in S2 is as follows: S2.

1. Use natural language processing NLP technology to detect ambiguities or errors in the user input; where: P(error∣input) represents the probability of error given the input, P(input∣error) represents the probability of the input in the case of error, P(error) represents the prior probability of error, and P(input) represents the prior probability of the input; S2.

2. Generate corresponding explanatory interaction content according to the detected error type; E = f(error_type, context) where: E represents the generated explanatory content, error_type represents the error type, and context represents the context information.

5. A method for alleviating robot conversation failure by using interpretive interaction according to claim 1, characterized in that, The method for providing personalized explanatory interaction content for each user by analyzing the user's interaction history and preference setting information in S3 is as follows: According to the user's preferences, select one or more of text, voice, image or video to output the explanatory interaction content; Among them: Output represents the final output explanatory content, w i represents the user's preference weight for the i-th form, E i represents the explanatory content of the i-th form.

6. The method for alleviating robot conversation failure by using interpretive interaction according to claim 1, wherein The method for optimizing the system according to the feedback information to continuously improve the quality and effect of the explanatory interaction in S4 is as follows: S4.

1. Adjust the robot conversation model according to user feedback and failure scenarios, and optimize the explanatory interaction strategy; S4.

2. Design a new explanatory interaction strategy for high-frequency failure scenarios and conduct A / B tests; redeploy the optimized system, continue to collect user feedback, and form a closed-loop optimization process; S4.

3. Compare the user satisfaction and session failure rate before and after optimization to evaluate the optimization effect; S4.

4. Establish a continuous monitoring mechanism, regularly evaluate the system performance, and ensure the continuous improvement of the quality and effect of explanatory interaction.

7. A method for alleviating robot conversation failure by using interpretive interaction according to claim 6, characterized in that, The method for optimizing the explanatory interaction strategy in S4.1 is as follows: Adjust the model parameter weights according to user satisfaction ω new = ω old + α × (Satisfaction score - Mean satisfaction score) where α is the learning rate, ω new represents the optimized interpretive interaction strategy parameter, ω old represents the interpretive interaction strategy parameter before optimization.

Images