Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12288031-B2
Publication Date
2025-04-29
Expiration Date
Abstract
Filtering user intents corresponding to user utterances is provided. A list of allowed user intents is generated, using a natural language understanding model of a chatbot, based on identifying one or more of a set of user intents corresponding to a user utterance within a filtered user intent mapping table. It is determined whether a user intent having a highest confidence score in the set of user intents corresponding to the user utterance is contained in the list of allowed user intents. In response to determining that the user intent having the highest confidence score in the set of user intents corresponding to the user utterance is contained in the list of allowed user intents, content corresponding to the user intent having the highest confidence score is sent, using the chatbot, to a client device of a user who submitted the user utterance as a response to the user utterance.
Core Innovation
The invention filters user intents corresponding to user utterances in a chatbot. A natural language understanding model of the chatbot identifies one or more of a set of user intents corresponding to a user utterance within a filtered user intent mapping table, and based on the identification the computer generates a list of allowed user intents corresponding to a user that submitted the user utterance.
The system determines whether a user intent having a highest confidence score in the set of user intents corresponding to the user utterance is contained in the list of allowed user intents. When the highest-confidence user intent is contained in the list, the system sends content corresponding to the highest-confidence user intent to a client device of the user as a response to the user utterance. When the highest-confidence user intent is not contained in the list of allowed user intents, the system determines that the user utterance is not understood or that the response is not allowed, and sends a clarification or not-allowed message to the client device of the user.
The filtered user intent mapping table is influenced by session initialization and user characteristics. The system receives a start-session input from the client device, analyzes session data to identify characteristics associated with the user, and initializes the user intent mapping table with user intent names and user intent eligibility filters based on the characteristics, including a valid user location filter. The system prunes intent names that do not match the valid user location filter and may also select additional user intent eligibility filters to conditionally remove user intent names that match identified user characteristics.
Claims Coverage
Independent claims are provided as a method, a computer system, a computer program product, and an additional method. Across these independents, the claims specify intent eligibility filtering using a filtered user intent mapping table, a membership check of the highest-confidence intent in an allowed-intent list, and response handling based on that check, with additional claim coverage for session-based eligibility filters and per-intent confidence score generation.
Allowed intent generation from a filtered user intent mapping table
Generating, by a computer using a natural language understanding model of a chatbot, a list of allowed user intents based on identifying one or more of a set of user intents corresponding to a user utterance within a filtered user intent mapping table, the filtered user intent mapping table containing allowed user intents corresponding to a user that submitted the user utterance to the chatbot.
Highest-confidence allowed-intent membership determination
Determining, by the computer, whether a user intent having a highest confidence score in the set of user intents corresponding to the user utterance is contained in the list of allowed user intents.
Sending response content for the highest-confidence allowed intent
Responsive to the computer determining that the user intent having the highest confidence score in the set of user intents corresponding to the user utterance is contained in the list of allowed user intents, sending, by the computer using the chatbot, content corresponding to the user intent having the highest confidence score to a client device of the user as a response to the user utterance.
Clarification or not-allowed messaging when the highest-confidence intent is not allowed
Responsive to determining that the user intent having the highest confidence score in the set of user intents corresponding to the user utterance is not contained in the list of allowed user intents, determining that the user utterance is not understood or the response is not allowed, and sending a message to the client device requesting clarification or stating the response is not allowed.
Confidence score generation for each user intent
Generating, by the computer using a natural language understanding model of a chatbot, a confidence score for each user intent in the set corresponding to a user utterance.
Session-based initialization of the user intent mapping table using user characteristics
Receiving a start-session input from the client device, analyzing session data to identify characteristics associated with the user, and initializing a user intent mapping table with intent names and eligibility filters based on those characteristics for the session.
Location-based pruning using a valid user location filter
Removing any user intent names from the user intent mapping table that do not match a valid user location filter from the set of user intent eligibility filters, using the user's current location from session data.
Conditional application of additional eligibility filters
Selecting an additional user intent eligibility filter from a set, determining whether a user characteristic in session data matches that selected filter, and if it matches removing user intent names from a user intent mapping table that correspond to the selected filter.
System-level implementation with bus, storage, and processor executing program instructions
A computer system for filtering user intents comprising a bus system, a storage device connected to the bus system storing program instructions, and a processor connected to the bus system executing the program instructions to generate the list of allowed user intents, determine whether the highest-confidence intent is contained in the list, and send content corresponding to the highest-confidence intent to a client device.
The independent claims cover a chatbot method and system that generate a list of allowed user intents from a filtered user intent mapping table, determine whether the highest-confidence intent is contained in that allowed list, and send corresponding response content when allowed. Additional dependent coverage specifies clarification or not-allowed messaging when not allowed, confidence score generation for intents, and session-based initialization and pruning of intent mapping entries using eligibility filters such as a valid user location filter and additional matching filters.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?