/* ---- Google Analytics Code Below */

Thursday, June 07, 2018

Learning from Process with Human Operators

Integration of human and AI learning processes.   A kind of focused crowdsourcing, where the operator scores machine learning and gets collaborative assistance from the machine based on context?   A good direction.

Self-Learning Assistance System for Efficient Processes   By Fraunhofer-Gesellschaft

The Fraunhofer Institute for Process Engineering and Packaging IVV in Germany is developing a self-learning assistance system called SAM to help machine operators correct errors while accruing experience and process knowledge. The SAM system monitors machine states and operator actions, retaining successful solution strategies. In one application, an operator with a tablet computer enters their solution and connects it to the current fault situation recorded by SAM. Should a given error happen multiple times, the system will recognize it and advise the operator on resolution strategies. The team is using machine learning algorithms to make SAM capable of learning fault scenarios, and thanks to intelligent feature extraction, the system's learning speed approaches that of humans. Fraunhofer IVV's Andre Schult says he wants SAM eventually to be utilized to enable machine operators to identify errors on their own, and to suggest their own remedies ... " 

No comments: