Back to results
US 11,775,772 B2Granted
Chatbot providing a defeating reply
AssigneeOracle International Corporation
InventorBoris Galitsky
Life of this patent
Estimating the term…
Abstract
Techniques are disclosed for generating defeating textual replies. A computer-implemented method generates a first communicative discourse tree from an utterance and as second communicative discourse tree from a candidate answer derived from a search of keywords in the utterance. The method translates the first communicative discourse tree into a first logical formula from and the second communicative discourse tree into a second logical formula. The method determines that the first logical formula is a negation of the second logical formula. The method further provides the candidate answer to a user device responsive to the determining.
Claims19
A method for negating a claim within an utterance, the method comprising:
- accessing, via a user device, a user utterance comprising fragments of text, wherein each fragment is an elementary discourse unit;
- generating, from the user utterance, a first discourse tree that represents rhetorical relationships between the fragments of text of the user utterance, wherein a discourse tree comprises a plurality of nodes, each nonterminal node representing a rhetorical relationship between two of the fragments, each terminal node of the nodes of the discourse tree being associated with one of the fragments;
- translating the first discourse tree into a first logical formula, wherein translating a discourse tree into a logical formula comprises:
- identifying, from the respective discourse tree, a rhetorical relation that corresponds to a nucleus elementary discourse unit and a satellite elementary discourse unit;
- mapping the nucleus elementary discourse unit to a reason and the satellite elementary discourse unit to a conclusion, wherein the mapping comprises:
- providing, to a trained machine learning model, the nucleus elementary discourse unit and the satellite elementary discourse unit, wherein the machine learning model is trained to predict whether an elementary discourse unit contains logical atoms associated with a reason or a conclusion, and
- receiving, from the trained machine learning model, the mapping;
- and
- constructing, from the identified rhetorical relation, the nucleus elementary discourse unit, and the satellite elementary discourse unit, a logical formula that includes the reason and the conclusion;
- identifying topic keywords from the user utterance;
- providing the topic keywords to a search engine;
- obtaining, from the search engine, a candidate answer;
- generating, from the candidate answer, a second discourse tree;
- translating the second discourse tree into a second logical formula;
- determining that the first logical formula is a negation of the second logical formula;
- and
- responsive to the determining, providing the candidate answer to the user device, wherein the user device presents the candidate answer as a response to the user utterance.
14 more elements
A non-transitory computer-readable storage medium storing computer-executable program instructions, wherein when executed by a processing device, the computer-executable program instructions cause the processing device to perform operations comprising:
- accessing, via a user device, a user utterance comprising fragments of text, wherein each fragment is an elementary discourse unit;
- generating, from the user utterance, a first discourse tree that represents rhetorical relationships between the fragments of text of the user utterance, wherein a discourse tree comprises a plurality of nodes, each nonterminal node representing a rhetorical relationship between two of the fragments, each terminal node of the nodes of the discourse tree being associated with one of the fragments;
- translating the first discourse tree into a first logical formula, wherein translating a discourse tree into a logical formula comprises:
- identifying, from the respective discourse tree, a rhetorical relation that corresponds to a nucleus elementary discourse unit and a satellite elementary discourse unit;
- mapping the nucleus elementary discourse unit to a reason and the satellite elementary discourse unit to a conclusion, wherein the mapping comprises:
- providing, to a trained machine learning model, the nucleus elementary discourse unit and the satellite elementary discourse unit, wherein the machine learning model is trained to predict whether an elementary discourse unit contains logical atoms associated with a reason or a conclusion, and
- receiving, from the trained machine learning model, the mapping;
- and
- constructing, from the identified rhetorical relation, the nucleus elementary discourse unit, and the satellite elementary discourse unit, a logical formula that includes the reason and the conclusion;
- identifying topic keywords from the user utterance;
- providing the topic keywords to a search engine;
- obtaining, from the search engine, a candidate answer;
- generating, from the candidate answer, a second discourse tree;
- translating the second discourse tree into a second logical formula;
- determining that the first logical formula is a negation of the second logical formula;
- and
- responsive to the determining, providing the candidate answer to the user device, wherein the user device presents the candidate answer as a response to the user utterance.
14 more elements
A system comprising:
- a non-transitory computer-readable medium storing computer-executable program instructions;
- and
- a processing device communicatively coupled to the non-transitory computer-readable medium for executing the computer-executable program instructions, wherein executing the computer-executable program instructions configures the processing device to perform operations comprising:
- accessing, via a user device, a user utterance comprising fragments of text, wherein each fragment is an elementary discourse unit;
- generating, from the user utterance, a first discourse tree that represents rhetorical relationships between the fragments of text of the user utterance, wherein a discourse tree comprises a plurality of nodes, each nonterminal node representing a rhetorical relationship between two of the fragments, each terminal node of the nodes of the discourse tree being associated with one of the fragments;
- translating the first discourse tree into a first logical formula, wherein translating a discourse tree into a logical formula comprises:
- identifying, from the respective discourse tree, a rhetorical relation that corresponds to a nucleus elementary discourse unit and a satellite elementary discourse unit;
- mapping the nucleus elementary discourse unit to a reason and the satellite elementary discourse unit to a conclusion, wherein the mapping comprises:
- providing, to a trained machine learning model, the nucleus elementary discourse unit and the satellite elementary discourse unit, wherein the machine learning model is trained to predict whether an elementary discourse unit contains logical atoms associated with a reason or a conclusion, and
- receiving, from the trained machine learning model, the mapping;
- and
- constructing, from the identified rhetorical relation, the nucleus elementary discourse unit, and the satellite elementary discourse unit, a logical formula that includes the reason and the conclusion;
- identifying topic keywords from the user utterance;
- providing the topic keywords to a search engine;
- obtaining, from the search engine, a candidate answer;
- generating, from the candidate answer, a second discourse tree;
- translating the second discourse tree into a second logical formula;
- determining that the first logical formula is a negation of the second logical formula;
- and
- responsive to the determining, providing the candidate answer to a user device, wherein the user device presents the candidate answer as a response to the user utterance.
17 more elements
Dependencies are inferred from the claim text.