Analog Computing for Artificial Intelligence

Analog integrated circuits, essential for AI, combine energy efficiency (100x improvement over digital) with parallel processing inspired by biological neural networks. This course presents cleanroom technologies for the manufacture of advanced transistors and non-volatile memories, as well as the design of circuits for analog computing (matrix multiplication, associative memory, analog-digital interfaces). By merging numerical precision with analog efficiency, these technologies overcome traditional computing limits, particularly in edge computing and neuromorphic AI. Students will master industrial processes and hybrid architectures, gaining the skills to design next-generation AI processors.

Introduction to Analog Computing for AI Principles of Microelectronics Fabrication Technologies Core Components:

• Matrix Multiplication: Crossbar Arrays
• Crossbar Structures: Physical Principles & Fabrication
• Neuromorphic Approaches: Spiking Neural Networks
• Analog-Digital Conversion Circuits

Hands-On Sessions: Lab: Introduction to cleanroom micro/nanotechnology Lab: Device Characterization Combined study/practical: Simulation of analog computing