As artificial intelligence’s (AI) sprawling influence reshapes industries from logistics and healthcare to automotive and manufacturing, Taiwan is poised to leverage its cutting-edge capabilities and rich history in semiconductor manufacturing to stake out a leadership position in AI. Taiwan’s semiconductor manufacturing industry accounts for a major share of the region’s GDP and, with its manufacturing prowess, the region is fertile ground for using AI to optimize and even revolutionize chip manufacturing.
In an AI and Semiconductor Smart Manufacturing Forum recently hosted by SEMI Taiwan, experts from Micronix, Advantech, Nvidia and the Ministry of Science and Technology of Taiwan (MOST) shared their insights on how deep learning, data analytics and edge computing will shape the future of semiconductor manufacturing. Here are four key takeaways.
1. Monitor, Forecast, and Prevent
Today, tier 1 foundries use AI tools to combine equipment know-how and manufacturing statistics in managing massive Fault Detection (FD) data, much in the way that a car’s tire-pressure monitoring system helps maintain safe inflation levels and prevent accidents. For example, AI enables the real-time collection and monitoring of massive amounts of processing data, then alerts system administrators of any hardware failures or other manufacturing abnormalities.
AI also makes it possible to adopt Run-to-Run (R2R) control to automate manufacturing process adjustments and corrections by providing feedback that can drive higher processing efficiency. In addition, virtual metrology replaces manual sampling inspection for comprehensive quality control, enabling foundries to improve yields, reduce costs, and strengthen their competitive advantage.
2. Beyond Automation: Edge Computing
How does edge computing work in a smart factory? First, a rich trove of data from various devices is collected and integrated via Manufacturing Execution Systems (MES). Software analysis then produces a real-time factory production status before production data is visualized through a combination of system platforms and human-machine interfaces. In the end, the data is analyzed realtime in the cloud so failures can be predicted and prevented to help increase capacity and reduce costs. The approach is even capable of Bill of Materials (BOM) predictions, allowing better collaboration between upstream and downstream suppliers.
3. Deep Learning Accelerates AI
Deep learning enables autonomous driving, intelligent voice assistance and many other AI breakthroughs. The heart of deep learning is its ability to automatically process and learn data in various formats such as images, video and text with no human domain knowledge. This increases predictive accuracy and efficiency in processing massive amounts of data. Deep learning also enhances the efficiency of human-machine collaboration.
4. Taiwan’s Competitive Niche: Industry 3.5
Industry 4.0 is not just about improving production management. It also focuses on integrating supply chains, even among competitive companies. For Industry 4.0 to thrive, rival companies must grow together. The first and third industrial revolutions centered on disruptive technologies like steam engines, transistors and digital, while the second and fourth revolutions homed in on competition among various business models, platforms and industry ecosystems.
While Taiwan’s strengths include innovation, short time-to-market, low manufacturing costs, and high supply chain management efficiency, the region still lags advanced countries in basic industry and research capabilities. Squeezed by Chinese supply chains and high-end manufacturers in advanced countries, Taiwan should start by carving out an Industry 3.5 niche for the island’s manufacturers.
SEMI will continue to facilitate cross-industry connection, collaboration and innovation to help manufacturers seeking higher production efficiency and lower costs incorporate AI as a core competitive advantage. At SEMICON Taiwan 2018, SEMI will unveil its Smart Manufacturing Journey, an exhibition that gathers leading AI companies such as ABB, Advantech, Nvidia, Sony and UPS to demonstrate a comprehensive roadmap for smart manufacturing technologies and applications. For more information, please visit the SEMICON Taiwan website.
Emmy Yi is a marketing specialist at SEMI Taiwan.