System and method for automated decision making

A non-transitory computer-readable medium storing instructions that, when executed by one or more computer processors of a computing platform, cause the computing platform to perform the operations:

• at a computing platform configured to communicate with multiple client devices, collecting a set of natural language response inputs to a problem prompt from the multiple client devices;
• at a similarity engine of the computing platform, consolidating, through processing of a machine learning model, the set of response inputs into a set of base responses, wherein consolidating the set of response inputs comprises:
• generating, using the machine learning model, similarity modeling across the set of response inputs,
• segmenting, based on the similarity modeling, the set of response inputs into responses groups,
• computationally determining a representative base response for each response group by processing response inputs of each response group with a predictive language model and outputting a generated base response;
• dynamically allocating pair-wise comparison tasks for comparisons of base responses to the multiple client devices;
• communicating the pair-wise comparison tasks to judgement interfaces of multiple client devices;
• receiving judgement input during a judgement stage in real-time from the multiple client devices, which comprises:
• through a response judgement interface at a client device, retrieving judgement input that includes a preference selection of one of the two base responses or a similarity selection of the two base responses;
• aggregating the received judgement input, wherein the computing platform monitors convergence of the aggregated preference rankings and automatically determines when to terminate the judgement stage based on the preference rankings reaching a stable state;
• wherein during the judgement stage, dynamically updating, based on the judgement input received from the multiple client devices, the set of base responses for pair-wise comparison which includes automatically re-consolidating by updating processing of the response inputs by the machine learning model, wherein judgment input indicating preference between two base responses is received at the machine learning model as attack scoring that trains the machine learning model to reduce similarity modeling between the two base responses and judgment input indicating similarity is received at the machine learning model as support scoring that trains the machine learning model to reinforce similarity modeling between the base responses, wherein the aggregated judgement input provides training data that continuously improves the machine learning model, and wherein the improved machine learning model is applied to subsequent dynamic allocation of pair-wise comparison tasks and subsequent consolidation of base responses across the multiple client devices;
• and
• generating a response report on preference ranking of a resulting set of base responses based on collected judgement input.

A system for distributed processing of natural language data across multiple client devices comprising:

• one or more computer-readable mediums storing instructions that, when executed by the one or more computer processors, cause a computing platform to perform operations comprising:
• at a computing platform in communication with multiple client devices, collecting a set of natural language response inputs to a problem prompt;
• at a similarity engine of the computing platform, consolidating, through processing of a machine learning model, the set of response inputs into a set of base responses, wherein consolidating the set of response inputs comprises:
• generating, using the machine learning model, similarity modeling across the set of response inputs,
• segmenting, based on the similarity modeling, the set of response inputs into responses groups,
• computationally determining a representative base response for each response group by processing response inputs of each response group with a predictive language model and outputting a generated base response;
• dynamically allocating pair-wise comparison tasks for comparisons of base responses to the multiple client devices;
• communicating the pair-wise comparison tasks to judgement interfaces of multiple client devices;
• receiving judgement input during a judgement stage in real-time from the multiple client devices, which comprises:
• through a response judgement interface at a client device, retrieving judgement input that includes a preference selection of one of the two base responses or a similarity selection of the two base responses;
• aggregating the received judgement input, wherein the computing platform monitors convergence of the aggregated preference rankings and automatically determines when to terminate the judgement stage based on the preference rankings reaching a stable state;
• wherein during the judgement stage, dynamically updating, based on the judgement input received from the multiple client devices, the set of base responses for pair-wise comparison which includes automatically re-consolidating by updating processing of the response inputs by the machine learning model, wherein judgment input indicating preference between two base responses is received at the machine learning model as attack scoring that trains the machine learning model to reduce similarity modeling between the two base responses and judgment input indicating similarity is received at the machine learning model as support scoring that trains the machine learning model to reinforce similarity modeling between the base responses, wherein the aggregated judgement input provides training data that continuously improves the machine learning model, and wherein the improved machine learning model is applied to subsequent dynamic allocation of pair-wise comparison tasks and subsequent consolidation of base responses across the multiple client devices;
• and
• generating a response report on preference ranking of a resulting set of base responses based on collected judgement input.