An overview of how Deep Learning is outperforming state of the art optimization techniques in all stages of the semiconductor manufacturing process. These disruptive technologies will drive improvements in capacity planning and prediction, cycle time reduction, fab throughput and equipment utilization, failure prediction and response etc. In aggregate this will make for a more robust and dynamic manufacturing process, which is much better equipped to respond to systemic shocks and disruptions.
While the promise of Deep Learning is very real, there remain challenges in applying these technologies at scale in today's fabs. We will conclude with a case study that addresses the challenges of deploying Intelligent Agents in a modern fab for improved scheduling and dispatching.