Collaborative decision making in dynamically changing environment

A method for collaborative decision making in a dynamically changing environment, the method comprises:

• receiving, by a processor, a query from a user using a graphical user interface (GUI), wherein the query corresponds to a problem requiring a decision;
• determining, by the processor, one or more intermediate steps required to reach the decision based on metadata associated to the problem, wherein the one or more intermediate steps are calculated using deep learning algorithms and artificial intelligence techniques;
• establishing, by the processor, a decision-making flow for the one or more intermediate steps required to reach the decision, wherein the decision-making flow corresponds to a sequence for execution of the one or more intermediate steps, and wherein the sequence is determined based on an importance and a severity of the one or more intermediate steps using deep learning algorithms;
• generating, by the processor, a decision space comprising one or more decision options, and wherein the one or more decision options are calculated based on the decision-making flow using artificial intelligence techniques and deep learning algorithms;
• crawling, by the processor, external data sources to automatically extract new information related to the decision space, wherein the external data sources comprise internet search, monitoring changes in published data repositories, subscription to market research services, and analysis of news and social media content;
• determining, by the processor, merit of the new information based on a set of factors comprising historic judgements, consequences, and credibility of the external data source, wherein the merit is determined using at least Artificial Intelligence (AI) techniques, Mathematical Modelling, Markov Decision Modelling, and Deep Learning (DL) algorithms on the set of factors;
• identifying, by the processor, one or more uncertain events based on the new information, wherein the one or more uncertain events cause an uncertain impact on the decision-making flow;
• calculating, by the processor, the uncertain impact caused by the one or more uncertain events using predictive algorithms and machine learning techniques, wherein the predictive algorithms are continuously trained based on reinforcement techniques;
• dynamically modifying, by the processor, the decision space based on the uncertain impact corresponding to the one or more uncertain events, wherein the decision space is modified by at least one of adding one or more decision options, and removing one or more decision options;
• generating, by the processor, a decision knowledge graph depicting each modification in the decision space based on the modified decision space;
• updating, by the processor, the decision space and the decision knowledge graph based on feedback, wherein the feedback is at least one of accepting and rejecting one or more decision options in the decision space;
• continuously training, by the processor, the predictive algorithms based on the feedback, wherein the predictive algorithms are continuously trained to dynamically modify the decision space based on the uncertain impact;
• and
• selecting, by the processor, the decision based on the updated decision knowledge graph and the updated decision space, wherein a final decision is selected post approval of the selected decision from the one or more human agents.

A non-transitory computer program product having embodied thereon a computer program for collaborative decision making in a dynamically changing environment, the computer program product storing instructions, the instructions comprising instructions for:

• receiving a query from a user using a graphical user interface (GUI), wherein the query corresponds to a problem requiring a decision;
• determining one or more intermediate steps required to reach the decision based on metadata associated to the problem, wherein the one or more intermediate steps are calculated using deep learning algorithms and artificial intelligence techniques;
• establishing a decision-making flow for the one or more intermediate steps required to reach the decision, wherein the decision-making flow corresponds to a sequence for execution of the one or more intermediate steps, and wherein the sequence is determined based on an importance and a severity of the one or more intermediate steps using deep learning algorithms;
• generating a decision space comprising one or more decision options, and wherein the one or more decision options are calculated based on the decision-making flow using artificial intelligence techniques and deep learning algorithms;
• crawling external data sources to automatically extract new information related to the decision space, wherein the external data sources comprise internet search, monitoring changes in published data repositories, subscription to market research services, and analysis of news and social media content;
• determining merit of the new information based on a set of factors comprising historic judgements, consequences, and credibility of the external data source, wherein the merit is determined using at least Artificial Intelligence (AI) techniques, Mathematical Modelling, Markov Decision Modelling, and Deep Learning (DL) algorithms on the set of factors;
• identifying one or more uncertain events based on the new information, wherein the one or more uncertain events cause an uncertain impact on the decision-making flow;
• calculating the uncertain impact caused by the one or more uncertain events using predictive algorithms and machine learning techniques, wherein the predictive algorithms are continuously trained based on reinforcement techniques;
• dynamically modifying the decision space based on the uncertain impact corresponding to the one or more uncertain events, wherein the decision space is modified by at least one of adding one or more decision options, and removing one or more decision options;
• generating a decision knowledge graph depicting each modification in the decision space based on the modified decision space;
• updating the decision space and the decision knowledge graph based on feedback, wherein the feedback is at least one of accepting and rejecting one or more decision options in the decision space;
• continuously training the predictive algorithms based on the feedback, wherein the predictive algorithms are continuously trained to dynamically modify the decision space based on the uncertain impact;
• and
• selecting the decision based on the updated decision knowledge graph and the updated decision space, wherein a final decision is selected post approval of the selected decision from the one or more human agents.