Kulicke & Soffa: Enabling predictive analytic through deep learning

The physical realization of SMART Manufacturing is decision making in real-world applications. The first section of the paper discusses the successful digital transformation journey of Kulicke & Soffa (K&S) through various critical initiatives. These include Value Stream Mapping (VSM) workshops, Manufacturing Execution System (MES), Data Warehouse cloud solution and finally Tableau visualization adoption. [1] In the second section, we will present the recent achievement of enabling predictive analytic through deep learning. In detail, we will discuss the challenges centering around 3 pillars of support; namely data infrastructure, data inference and advanced analytics. Data infrastructure translates the real-world manufacturing processes into the digital world. Data inference links up the data generation to the business case. This iteratively refines data infrastructure, product design and manufacturing design for maximum exploitation. Lastly, suitable advanced analytic techniques must be chosen to optimize the defined system for performance improvement. Being successful in these 3 pillars revolves around an iterative development of synergies between physical sciences subject matter experts (SMEs), manufacturing operations and the analytics team.