Tools and Methodologies for Edge-AI Mixed-Signal Inference Accelerators

Join senior engineers Maen Mallah and Roland Müller from the Fraunhofer Institute for Integrated Circuits IIS in a one-hour technical talk exploring innovative approaches to edge-AI acceleration. Learn about the development of neuromorphic mixed-signal edge-AI accelerators and the challenges of hardware/software co-design. Discover specialized tools designed to facilitate the design, training, and deployment of artificial neural networks in dedicated hardware accelerators, including hardware-aware training, automatic hardware generation, compilers, and KPI estimation. Explore how this multidisciplinary toolchain optimizes hardware design, achieves target performance metrics, and accelerates time-to-market while addressing the computational challenges of implementing neural networks at the edge and near sensors.