Augmenting an answer set
AssigneeMocha Technologies Inc.
InventorsSujit Karpe, VISHAL MADAN, WOLFGANG HOLTGEN
Life of this patent
Estimating the term…
Abstract
A method and a system for augmenting an answer set on an online platform is disclosed. The method includes receiving an answer to a question. The answer may be received from a plurality of candidates. The method further includes comparing the answer with a predefined answer. The method furthermore includes identifying a deviation between the answer and the predefined answer. The method furthermore includes counting a number of occurrences of the answer. The method furthermore includes alerting when the number of occurrences of the answer exceeds a predefined threshold. Finally, the method includes augmenting a predefined answer set by adding the answer post validating the answer.
Claims13
A method for augmenting an answer set on an online platform, the method comprises:
- receiving, by a processor, an answer to a question, wherein the answer is received from a plurality of candidates;
- comparing, by the processor, the answer with a predefined answer using Machine Learning Techniques, Natural Language Processing (NLP), and Text Similarity Techniques;
- identifying, by the processor, a deviated answer, wherein the deviated answer is the answer deviated from the predefined answer, and wherein deviation between the answer and the predefined answer is identified based upon the comparison, and wherein the deviation is identified based upon vectorization, cosine similarity, Euclidean distance, and machine learning based text similarity techniques;
- counting, by the processor, a number of occurrences of the deviated answer;
- alerting, by the processor, when the number of occurrences of the deviated answer exceeds a predefined threshold;
- and
- augmenting, by the processor, a predefined answer set by adding the deviated answer post validating based on the alert, wherein the predefined answer set is augmented by using a supervised learning approach.
(canceled)
A system for augmenting an answer set on an online platform, the system comprising:
- a memory;
- and
- a processor coupled to the memory, wherein the processor is configured for:
- receiving an answer to a question, wherein the answer is received from a plurality of candidates;
- comparing the answer with a predefined answer using Machine Learning Techniques, Natural Language Processing (NLP), and Text Similarity Techniques;
- identifying a deviated answer, wherein the deviated answer is the answer deviated from the predefined answer, and wherein deviation between the answer and the predefined answer is identified based upon the comparison, and wherein the deviation is identified based upon vectorization, cosine similarity, Euclidean distance, and machine learning based text similarity techniques;
- counting a number of occurrences of the deviated answer;
- alerting when the number of occurrences of the deviated answer exceeds a predefined threshold;
- and
- augmenting, based on the alert, a predefined answer set by adding the deviated answer post validating, wherein the predefined answer set is augmented by using a supervised learning approach.
(canceled)
A non-transitory computer program product having embodied thereon a computer program for augmenting an answer set on an online platform, the computer program product storing instructions, the instructions for:
- receiving an answer to a question, wherein the answer is received from a plurality of candidates;
- comparing the answer with a predefined answer using Machine Learning Techniques, Natural Language Processing (NLP), and Text Similarity Techniques;
- identifying a deviated answer, wherein the deviated answer is the answer deviated from the predefined answer, and wherein deviation between the answer and the predefined answer is identified based upon the comparison, and wherein the deviation is identified based upon vectorization, cosine similarity, Euclidean distance, and machine learning based text similarity techniques;
- counting a number of occurrences of the deviated answer;
- alerting when the number of occurrences of the deviated answer exceeds a predefined threshold;
- and
- augmenting, based on the alert, a predefined answer set by adding the answer post validating, wherein the predefined answer set is augmented by using a supervised learning approach.
Dependencies are inferred from the claim text.