
Brains and Machines Choosing the Right Technologies for Hybrid AI Chips
Apr 5, 2024
Amirreza Yousefzadeh, a neuromorphic hardware architect with a rich background in designing scalable hybrid AI systems on a chip, discusses the advancements in neuromorphic technology. He shares insights on balancing research and commercialization, particularly with the development of the Seneca chip. The conversation delves into the integration of neuromorphic systems in automotive applications, the need for better market demonstrations, and challenges in traditional machine vision. Plus, there's a spotlight on the role of power efficiency in driving innovation.
AI Snips
Chapters
Transcript
Episode notes
PhD Journey
- Amirreza Yousefzadeh's PhD journey began with the Human Brain Project and Spinnaker.
- Early excitement turned to disillusionment, then a focus on market value with BrainChip.
Market Focus
- Yousefzadeh's PhD research at IMEC explored event-based vision and neuromorphic chips.
- Collaboration with BrainChip shifted his focus to market-driven product development.
Gray Matter and IMEC
- Yousefzadeh left his postdoc to become a hardware architect at Gray Matter Labs.
- He aimed for a hybrid processor for both compute and neural networks, balancing research and market demands.
