Dialogue system, a dialogue method and a method of adapting a dialogue system
First Claim
1. A dialogue system comprising:
- an input for receiving data relating to a speech or text signal originating from a user;
an output for outputting information specified by an action; and
a processor configured to;
i) update a system state based on the input data using a state tracker model, wherein the system state comprises probability values associated with each of a plurality of possible values for each of a plurality of categories, wherein a category corresponds to a subject that the speech or text signal may relate to and can take on one or more values from a set of values;
ii) identify one or more relevant categories based on at least part of the updated system state information using an identifier model;
iii) define a set of information from stored information, the stored information comprising a plurality of action functions and categories, the set of information excluding categories which were not identified as relevant;
iv) generate a reduced system state comprising the probability values associated with one or more of the plurality of possible values for each of the relevant categories;
v) determine an action based on the reduced system state and the set of information using a policy model; and
vi) output information specified by the determined action at the output, wherein the action comprises an action function combined with a relevant category or an action function,wherein the stored information and the plurality of categories are from a plurality of different pre-defined domain-specific ontologies, wherein a pre-defined domain corresponds to a particular dialogue topic and wherein the pre-defined domain-specific ontology comprises the plurality of categories and actions relating to the particular dialogue topic, and wherein the relevant categories comprise categories from different pre-defined domain-specific ontologies and the set of information comprises actions from different pre-defined domain-specific ontologies.
1 Assignment
0 Petitions
Accused Products
Abstract
A dialogue system including: an input receiving data relating to a speech or text signal originating from a user; and a processor configured to: update a system state based on the input data using a state tracker model, the system state including probability values associated with each of plural possible values for each of plural categories; identify one or more relevant categories based on at least part of the updated system state information using an identifier model; define a set of information from stored information including plural action functions and categories, excluding categories not identified as relevant; generate a reduced system state, including the probability values associated with one or more of the plural possible values for each relevant category; determine an action based on the reduced system state and the set of information using a policy model; output information specified by the determined action at an output.
9 Citations
17 Claims
-
1. A dialogue system comprising:
-
an input for receiving data relating to a speech or text signal originating from a user; an output for outputting information specified by an action; and a processor configured to; i) update a system state based on the input data using a state tracker model, wherein the system state comprises probability values associated with each of a plurality of possible values for each of a plurality of categories, wherein a category corresponds to a subject that the speech or text signal may relate to and can take on one or more values from a set of values; ii) identify one or more relevant categories based on at least part of the updated system state information using an identifier model; iii) define a set of information from stored information, the stored information comprising a plurality of action functions and categories, the set of information excluding categories which were not identified as relevant; iv) generate a reduced system state comprising the probability values associated with one or more of the plurality of possible values for each of the relevant categories; v) determine an action based on the reduced system state and the set of information using a policy model; and vi) output information specified by the determined action at the output, wherein the action comprises an action function combined with a relevant category or an action function, wherein the stored information and the plurality of categories are from a plurality of different pre-defined domain-specific ontologies, wherein a pre-defined domain corresponds to a particular dialogue topic and wherein the pre-defined domain-specific ontology comprises the plurality of categories and actions relating to the particular dialogue topic, and wherein the relevant categories comprise categories from different pre-defined domain-specific ontologies and the set of information comprises actions from different pre-defined domain-specific ontologies. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10)
-
-
11. A dialogue method comprising:
-
receiving, in an input, data relating to a speech or text signal originating from a user; updating, in a processor, a system state based on the input data using a state tracker model, wherein the system state comprises probability values associated with each of a plurality of possible values for each of a plurality of categories, wherein a category corresponds to a subject that the speech or text signal may relate to and can take on one or more values from a set of values; identifying, in the processor, one or more relevant categories based on at least part of the updated system state information using an identifier model; defining, in the processor, a set of information from stored information, the stored information comprising a plurality of action functions and categories, the set of information excluding categories which were not identified as relevant; generating, in the processor, a reduced system state, comprising the probability values associated with one or more of the plurality of possible values for each of the relevant categories; determining, in the processor, an action based on the reduced system state and the set of information using a policy model; and outputting, in an output, information specified by the determined action at the output, wherein the action comprises an action function combined with a relevant category or an action function, wherein the stored information and the plurality of categories are from a plurality of different pre-defined domain-specific ontologies, wherein a pre-defined domain corresponds to a particular dialogue topic and wherein the pre-defined domain-specific ontology comprises the plurality of categories and actions relating to the particular dialogue topic, and wherein the relevant categories comprise categories from different pre-defined domain-specific ontologies and the set of information comprises actions from different pre-defined domain-specific ontologies.
-
-
12. A method of adapting a dialogue system, the method comprising:
-
obtaining, in a processor, system states for a plurality of dialogues, each system state based on a speech or text signal in a dialogue, wherein the system state comprises probability values associated with each of a plurality of possible values for each of a plurality of categories, wherein a category corresponds to a subject that the speech or text signal may relate to and can take on one or more values from a set of values; and training, in the processor, an identifier model to identify one or more relevant categories based on at least part of an input system state information, wherein the identifier model and a policy model are trained together by using the identifier model and the policy model to run through dialogues with a human or a simulated human and providing a performance indicator for each dialogue, and the method further comprises; receiving, in an input, data relating to a speech or text signal in a dialogue, wherein obtaining a system state comprises updating a system state based on the input data using a state tracker model, identifying, in the processor, one or more relevant categories based on at least part of the updated system state information using the identifier model; defining, in the processor, a set of information from stored information, the stored information comprising a plurality of action functions and categories, the set of information excluding categories which were not identified as relevant; generating, in the processor, a reduced system state, comprising the probability values associated with one or more of the plurality of possible values for each of the relevant categories; determining, in the processor, an action based on the reduced system state and the set of information using a policy model; outputting, in an output, information specified by the determined action at the output; and adapting the policy model and the identifier model to increase the performance indicator. - View Dependent Claims (13, 14, 15, 16)
-
-
17. A non-transitory carrier medium comprising computer readable code that, when executed by a computer, causes the computer to perform:
-
receiving data relating to a speech or text signal originating from a user; updating a system state based on the input data using a state tracker model, wherein the system state comprises probability values associated with each of a plurality of possible values for each of a plurality of categories, wherein a category corresponds to a subject that the speech or text signal may relate to and can take on one or more values from a set of values; identifying one or more relevant categories based on at least part of the updated system state information using an identifier model; defining a set of information from stored information, the stored information comprising a plurality of action functions and categories, the set of information excluding categories which were not identified as relevant; generating a reduced system state, comprising the probability values associated with one or more of the plurality of possible values for each of the relevant categories; determining an action based on the reduced system state and the set of information using a policy model; and outputting information specified by the determined action at the output, wherein the action comprises;
an action function combined with a relevant category;
or an action function,wherein the stored information and the plurality of categories are from a plurality of different pre-defined domain-specific ontologies, wherein a pre-defined domain corresponds to a particular dialogue topic and wherein the pre-defined domain-specific ontology comprises the plurality of categories and actions relating to the particular dialogue topic, and wherein the relevant categories comprise categories from different pre-defined domain-specific ontologies and the set of information comprises actions from different pre-defined domain-specific ontologies.
-
Specification