I am working on contextual intelligent automation (ML Model) for invoice approval. I am using Decision workbenck to capture human decision based on model prediction score.
My expectation - ML model should learn human decision (SME) and take corresponding actions in next similar request but expected behaviour is not working.
Please help me to acheive this scenario.
AssistEdge Enterprise 19.1 - Decision Workbench provides capability to capture human decisions based on model prediction score, self learning mechanisms based on human decisions where confidence score is calculated - if that is more than configured acceptable limit, automation process workflow sends status to the decision workbench application, then the Subject Matter Expert takes further actions - approvals.
You can refer AssistEdge Enterprise 19.1 Knowledge base - a sample use case can be reused to achieve the mentioned scenario -
Thanks @sumit.sagar for your response.
But my requirement is, how to retraining ML model based on human response? Please help me.
You need to create a feedback loop for your ML model which will train your ML model depending upon inputs recieved by human user. Out of box feature for creating a feedback loop from AE RPA system to ML model is currently not present