Human-computer interaction method and electronic device
AssigneeHuawei Technologies Co., Ltd.
InventorsQing Zhang, Jinhui Zhang, Yibo Zhang
Life of this patent
Estimating the term…
Abstract
A method specifically includes, in a human-computer dialog interaction process, when a server performs slot extraction on user discourse, if there is a slot corresponding to which information fails to be extracted and the slot is an optional key slot, the server asks the user a question to determine whether the information corresponding to the slot is necessary. If the information is necessary, the server further extracts the information corresponding to the slot. If the information is unnecessary, the server does not extract the information corresponding to the slot.
Claims20
A human-computer interaction method, implemented by a human-computer dialog system, wherein the human-computer interaction method comprises:
- receiving a first input that comprises a service requirement of a user;
- determining, based on the first input, a first field corresponding to the first input, wherein the first field is a task scenario corresponding to the service requirement;
- distributing the first input to an intention recognition model corresponding to the first field;
- recognizing a first intention corresponding to the first input and based on the intention recognition model, wherein the first intention is a sub-scenario of the first field;
- attempting to extract, from the first input, first information corresponding to a first slot in the first intention, wherein the first slot is preconfigured in the first intention, and wherein the first slot is an optional key slot;
- asking, when the first information fails to be extracted, the user a first question to determine whether the first information is necessary;
- receiving a second input that comprises second information from the user to determine whether the first information is necessary;
- when the second input indicates that the first information, is necessary, extracting the first information from the second input and performing, based on the first intention and the first information, an operation corresponding to the first intention, and
- when the second input indicates that the first information is unnecessary, skipping extracting the first information and performing, based on the first intention, the operation.
A server, comprising:
- a processor;
- and
- a memory coupled to the processor and configured to store instructions that, when executed by the processor, cause the server to be configured to:
- receive a first input that comprises a service requirement of a user;
- determine, based on the first input, a first field corresponding to the first input, wherein the first field is a task scenario corresponding to the service requirement;
- distribute the first input to an intention recognition model corresponding to the first field;
- recognize a first intention corresponding to the first input and based on the intention recognition model, wherein the first intention is a sub-scenario of the first field;
- attempt to extract, from the first input, first information corresponding to a first slot in the first intention, wherein the first slot is preconfigured in the first intention, and wherein the first slot is an optional key slot;
- ask, when the first information fails to be extracted, the user a question to determine whether the first information is necessary;
- receive a second input that comprises second information from the user to determine whether the first information is necessary;
- when the second input indicates that the first information is necessary, extract the first information from the second input and perform, based on the first intention and the first information, an operation corresponding to the first intention;
- and
- when the second input indicates that the first information is unnecessary, skip extracting the first information and perform, based on the first intention, the operation.
A computer program product comprising computer-executable instructions stored on a non-transitory computer-readable medium that, when executed by a processor, cause a human-computer dialog system to:
- receive a first input that comprises a service requirement of a user;
- determine, based on the first input, a first field corresponding to the first input, wherein the first field is a task scenario corresponding to the service requirement;
- distribute the first input to an intention recognition model corresponding to the first field;
- recognize a first intention corresponding to the first input and based on the intention recognition model, wherein the first intention is a sub-scenario of the first field;
- attempt to extract, from the first input, first information corresponding to a first slot in the first intention, wherein the first slot is preconfigured in the first intention, and wherein the first slot is an optional key slot;
- ask, when the first information fails to be extracted, the user a first question to determine whether the first information is necessary;
- receive a second input that comprises second information from the user to determine whether the first information is necessary;
- when the second input indicates that the first information is necessary, extract the first information from the second input and perform, based on the first intention and the first information, an operation corresponding to the first intention;
- and
- when the second input indicates that the first information is unnecessary, skip extracting the first information and perform, based on the first intention, the operation.
Dependencies are inferred from the claim text.