| CPC G06F 40/35 (2020.01) [G06F 16/2457 (2019.01); G06F 16/24575 (2019.01); G06F 16/24578 (2019.01); G06F 16/901 (2019.01); G06F 16/9035 (2019.01); G06F 16/90332 (2019.01); G06F 40/56 (2020.01); G06N 3/08 (2013.01); G06N 5/04 (2013.01)] | 19 Claims |
|
1. A system comprising:
a processor; and
memory storing executable instructions, which, when executed by the processor cause the system to:
obtain user information through a current interaction between a virtual agent associated with the system and a user via one or more communication channels, the virtual agent being a digital assistant with automated chat interfaces;
determine a plurality of utterance influencing attributes influencing a response recommended for user query from a plurality of sources, wherein the plurality of utterance influencing attributes are determined based on the obtained information, a user profile, and data related to external factors;
select a set of utterance influencing attributes among the plurality of utterance influencing attributes based on correlations among the plurality of utterance influencing attributes and collate the selected set of utterance influencing attributes to provide enriched user data;
identify a set of preconfigured potential utterance options associated with the user query;
implement a plurality of utterance recommendation techniques comprising machine learning techniques, deep learning techniques, orthogonal vector based techniques, and collaborative filtering techniques to analyze the enriched user data;
analyze the enriched data based on predefined rules associated with the utterance recommendation technique;
based on the analysis, provide a preferential rank ordering of the preconfigured potential utterance options to be recommended;
receive a plurality of preferential rank orderings of the preconfigured potential utterance options, the plurality of preferential rank orderings including the preferential rank ordering of the preconfigured potential utterance options provided by each of the plurality of the utterance recommendation techniques;
determine a single optimized rank ordering of the preconfigured potential utterance options by aggregating the plurality of preferential rank orderings provided by the plurality of the utterance recommendation techniques, wherein the aggregation is performed using parallel consensus aggregation with a minimum final dissimilarity in the plurality of preferential rank orderings, the single optimized rank ordering representing a ranking order of the preconfigured potential utterance options; and
enable the virtual agent to mimic human-like interactions using a potential utterance option ranked highest in the single optimized rank ordering, thereby facilitating dynamic human-like interaction between the virtual agent and the user.
|