Using classified text or images and deep learning algorithms to identify risk of product defect and provide early warning

摘要

Deep learning is used to identify specific, potential risks to an enterprise (of which product liability is the prime example here) while such risks are still internal electronic communications. The system involves mining and using existing classifications of data (e.g., from an internal litigation database, or from external sources such as customer complaints, and/or warranty claims) to train one or more deep learning algorithms, and then examining the enterprise's internal electronic communications with the trained algorithm, to generate a scored output that will enable enterprise personnel to be alerted to risks and take action in time to prevent the risks from resulting in harm to the enterprise or others.