Claims:WE CLAIM:
1. A method of processing multimodal user queries 105, the method comprising:
determining, by a response generation system 101, availability of one or more responses 110 to each of one or more sub-queries 106, wherein the one or more sub-queries 106 are formed by splitting the multimodal user queries 105;
detecting, by the response generation system 101, requirement of an expert 109 to provide the one or more responses 110 upon determining at least one of unavailability of the one or more responses 110 by the response generation system 101 or based on a first pre-defined condition;
generating, by the response generation system 101, a summarized content 107 by summarizing context of the one or more sub-queries 106 and historical conversation data associated with the one or more sub-queries 106;
reformulating, by the response generation system 101, the one or more sub-queries 106 based on the summarized content 107; and
collating, by the response generation system 101, the one or more responses 110 received, from the expert 109, for the reformulated one or more sub-queries 108.
2. The method as claimed in claim 1 further comprises providing, by the response generation system 101, the one or more responses 110 to the one or more sub-queries 106 upon detecting the availability of the one or more responses 110 for the one or more sub-queries 106.

3. The method as claimed in claim 1, wherein determining the availability of the one or more responses 110 to the one or more sub-queries 106 comprises:
creating, dynamically, by the response generation system 101, an input frame comprising one or more input slots, wherein a count of the one or more input slots created depends on a count of the one or more sub-queries 106; and
detecting, by the response generation system 101, the availability of the one or more responses 110 to the one or more sub-queries 106 when each of the one or more input slots are occupied with respective one of the one or more responses 110 and when weighted sum of each of the one or more responses 110 exceed a predefined threshold value.
4. The method as claimed in claim 3, wherein the first pre-defined condition comprises:
detection of an ambiguity in the one or more sub-queries 106 based on a vacancy in the one or more input slots, or
unavailability of a unique response to the one or more input slots created based on the one or more sub-queries 106.
5. The method as claimed in claim 3, wherein the weighted sum is obtained based on a second pre-defined condition comprising a degree of matching a keyword in each of the one or more sub-queries 106 with the one or more responses 110 provided by the response generation system 101, a confidence level in each of the one or more responses 110, or a count of responses 110 provided by the response generation system 101.

6. The method as claimed in claim 1, wherein collating each of the one or more responses 110 comprises:
creating, dynamically, by the response generation system 101, an output frame comprising one or more output slots, wherein a count of the one or more output slots created depends on a count of the reformulated one or more sub-queries 108;
detecting, by the response generation system 101, availability of the one or more responses 110 for each of the reformulated one or more sub-queries 108, from the expert 109, when each of the one or more output slots are occupied with one of the one or more responses 110; and
collating, by the response generation system 101, each of the one or more responses 110 upon detecting the availability of the one or more responses 110 for each of the reformulated one or more sub-queries 108.
7. The method as claimed in claim 6 further comprises detecting, by the response generation system 101, unavailability of the one or more responses 110, from the expert 109, for the reformulated one or more sub-queries 108 when the one or more output slots are unoccupied with the one of the one or more responses 110.

8. The method as claimed in claim 1 further comprises querying a user for receiving additional information upon identifying insufficiency in the reformulated one or more sub-queries 108 for providing the one or more responses 110.

9. The method as claimed in claim 1, wherein the collating each of the one or more responses 110 comprises removing duplicate sentences in each of the one or more responses 110.

10. The method as claimed in claim 1, wherein the multimodal user queries 105 comprises user queries provided in one or more input formats comprising text, audio and gesture.

11. A response generation system 101 for processing multimodal user queries 105, the system comprising:
a processor 203; and
a memory, communicatively coupled to the processor 203, wherein the memory stores processor-executable instructions, which, on execution, causes the processor 203 to:
determine availability of one or more responses 110 to each of one or more sub-queries 106, wherein the one or more sub-queries 106 are formed by splitting the multimodal user queries 105;
detect requirement of an expert 109 to generate the one or more responses 110 upon determining at least one of unavailability of the one or more responses 110 by the response generation system or based on a first predefined condition;
generate a summarized content 107 by summarizing context of the one or more sub-queries 106 and historical conversation data associated with the one or more sub-queries 106;
reformulate the one or more sub-queries 106 based on the summarized content 107; and
collate the one or more responses 110 received, from the expert 109, for the reformulated one or more sub-queries 108.
12. The response generation system 101 as claimed in claim 11, wherein the processor 203 provides the one or more responses 110 to the one or more sub-queries 106 upon detecting the availability of the one or more responses 110 for the one or more sub-queries 106.

13. The response generation system 101 as claimed in claim 11, wherein, to determine the availability of the one or more responses 110 to the one or more sub-queries 106, the processor 203:
creates dynamically an input frame comprising one or more input slots, wherein a count of the one or more input slots created depends on a count of the one or more sub-queries 106; and
detects the availability of the one or more responses 110 to the one or more sub-queries 106 when each of the one or more input slots are occupied with respective one of the one or more responses 110 and when weighted sum of each of the one or more responses 110 match a predefined threshold value.
14. The response generation system 101 as claimed in claim 13, wherein a first pre-defined condition comprises:
detection of an ambiguity in the one or more sub-queries 106 based on a vacancy in the one or more input slots, or
unavailability of a unique response to the one or more input slots created based on the one or more sub-queries 106.
15. The response generation system 101 as claimed in claim 13, wherein the weighted sum is obtained based on a second predefined condition comprising a degree of matching a keyword in each of the one or more sub-queries 106 with the one or more responses 110 provided by the response generation system, a confidence level in each of the one or more responses 110 or a count of responses provided by the response generation system 101.

16. The response generation system 101 as claimed in claim 12, wherein to collate each of the one or more responses 110, the processor 203:
creates dynamically an output frame comprising one or more output slots, wherein a count of the one or more output slots created depends on a count of the reformulated one or more sub-queries 108;
detects availability of the one or more responses 110 for each of the reformulated one or more sub-queries 108, from the expert 109, when each of the one or more output slots are occupied with one of the one or more responses 110; and
collates each of the one or more responses 110 upon detecting the availability of the one or more responses 110 for each of the reformulated one or more sub-queries 108.
17. The response generation system 101 as claimed in claim 18, wherein the processor 203 detects unavailability of the one or more responses 110, from the expert 109, for the reformulated one or more sub-queries 108 when the one or more output slots are unoccupied with the one of the one or more responses 110.

18. The response generation system 101 as claimed in claim 12, wherein the processor 203 queries a user for receiving additional information upon identifying insufficiency in the reformulated one or more sub-queries 108 for providing the one or more responses 110.

Dated this 24th day of November 2017

SWETHA S. N
OF K&S PARTNERS
ATTORNEY FOR THE APPLICANT
, Description:TECHNICAL FIELD
The present subject matter is generally related to artificial intelligence and more particularly, but not exclusively, to a method and system for processing multimodal user queries for generating automated responses.