Brains and Machines
Dr Sunny Bains
Curious to explore the technology advancing Artificial Intelligence beyond the usual headlines? Brains and Machines will introduce you to the people and ideas behind neuromorphic engineering, bio-inspired robotics, and other transformative technologies shaping AI’s future. From spiking neural networks and event-cameras to models of attention and mechanisms for prosthetic control, we investigate how machine cognition is moving forward.Join Dr Sunny Bains, a scientist, journalist, and lecturer at University College London, as she talks to researchers, engineers, and computer scientists from across the field. With co-host, Dr Giulia D’Angelo from the Czech Technical University in Prague, and commentator Prof Ralph Etienne-Cummings from Johns Hopkins University in Baltimore, the post-interview discussion provides context and insight into the featured innovations.Produced in conjunction with Electronic Engineering Times. Check out the EETimes Current podcast for more.Dr D’Angelo gratefully acknowledges the support of the Marie Skłodowska-Curie Postdoctoral Fellowship.
Episodes
Mentioned books
Feb 16, 2024 • 44min
Algorithms Help Spiking Neural Networks Learn to Learn
Emre Neftci, director at the Peter Grünberg Institute, dives into the fascinating world of spiking neural networks. He reveals how his research is transforming backpropagation for more efficient learning, mirroring brain functions. The conversation highlights the potential of neuromorphic computing in robotics and AI memory management. Neftci discusses innovations in prosthetic technologies, showcasing the exceptional capabilities of interfacing nerve fascicles. This insightful dialogue explores the future of intelligent systems inspired by our own biology.
Feb 9, 2024 • 45min
Giving the Humanoid iCub Embodied, Neuromorphic Vision
Dr. Chiara Bartolozzi, head of the event-driven perception for the robotics group at the Italian Institute for Technology (IIT) in Genoa, develops analog sub-threshold circuits to make bio-inspired brains for robots. Her group focuses on exploiting information from event-driven vision and tactile sensors for cognitive tasks, and she works extensively with iCub: a research platform in the form of a robot child, developed in Italy and used throughout Europe.In this episode of Brains and Machines, you’ll hear her talk to Dr. Sunny Bains about how neuromorphic technology can be used to implement attention mechanisms, the importance of embodiment, and why we need a solid theory of how neural systems can work together to create intelligence.Discussion follows with Dr. Giulia D’Angelo from the Czech Technical University in Prague, and Professor Ralph Etienne-Cummings of Johns Hopkins University.
Jan 30, 2024 • 44min
Tobi Delbrück Talks Caltech, Cameras, and Neural Control
Tobi Delbrück, a pioneer in neuromorphic engineering and professor at the Institute of Neuroinformatics in Zurich, delves into his journey from Caltech to cutting-edge technology. He discusses the evolution of dynamic vision sensors, highlighting how they can mimic human sight for better efficiency. The conversation also tackles the role of sparsity in enhancing neural hardware, impacting fields like robotics and speech recognition. Delbrück emphasizes the shift from theory to practice, showcasing the real-world applications of neuromorphic computing.
Jan 12, 2024 • 40min
Ferroelectric Memristors and Exotic Materials to Drive AI
In this episode of Brains and Machines, Sunny Bains gets deep into nanoscale ferroelectrics with Professor Beatriz Noheda, director of the Groningen Cognitive Systems and Materials Center (CogniGron). They discuss how this unusual interdisciplinary research center works, why nanoscale ferroelectrics may be useful in brain-like systems, and a little about how they are designed and fabricated.
Dec 15, 2023 • 43min
Perceiving Touch With Event-Based Neuromorphic Computing
Simeon Bamford, a senior researcher at the Italian Institute of Technology, specializes in tactile neuromorphic sensors and the commercialization of dynamic vision sensors. He discusses how circuits might restore lost functions due to brain damage. The conversation also covers advancements in tactile sensors that enhance robotic touch, contrasting traditional and event-based technologies. Bamford highlights the potential of his work in treating neurological conditions, emphasizing the importance of practical applications in neuroscience and robotics.
Dec 12, 2023 • 48min
How Memristors Will Help Machines Think at Different Timescales
Malika Pavant, Head of the Emergent Intelligence Substrates Lab in Zurich, delves into the fascinating world of memristors and their transformative potential for machines. She explains how these devices enhance neural systems' efficiency by mimicking biological features. The conversation covers the importance of adaptability in autonomous robots and discusses new strategies for neural networks to function across varying timescales. Pavant highlights the synergy of resistive memory technologies with traditional silicon, paving the way for more intelligent machines.
Nov 17, 2023 • 43min
ARM Inventor Steve Furber on SpiNNaker 1, 2 and Beyond
Steve Furber, Emeritus Professor at the University of Manchester and co-designer of the ARM architecture, shares his groundbreaking work in neuromorphic engineering. He discusses the evolution of associative memories and the design innovations behind the SpiNNaker platform, emphasizing brain-inspired computing's potential. Furber also explores challenges in timing and communication within neuromorphic systems and hints at the advancements coming with SpiNNaker 2, paving the way for improved AI applications and robotic navigation.
Oct 27, 2023 • 44min
How to See ‘Where’ Through Low-Power Event Cameras
Professor Guillermo Gallego, an expert in algorithms for event-based vision from the Technical University of Berlin, discusses the cutting-edge world of bio-inspired event-driven cameras. He explains how these cameras mimic human vision and operate in various light conditions, which is vital for autonomous robotics. The conversation dives into motion estimation, enhancing spatial AI, and the unifying frameworks that improve visual processing. Gallego also shares insights on the historical evolution of cellular neural networks and their role in modern imaging systems.
Oct 13, 2023 • 46min
Seeing Robotics and Machine Vision as Dynamical Systems
Yulia Sandamirskaya, creator of the Neuromorphic Computing Group at Zurich University of Applied Sciences, combines her expertise in dynamical systems theory and robotics. She dives into the intersection of neural networks and robotics, highlighting innovative applications like drone demos and insect-inspired designs. The discussion explores the challenges of robotic place recognition and the influence of biological principles on robot design. Sandamirskaya emphasizes the importance of 3D spatial representation and active perception for advancing AI capabilities.
Oct 6, 2023 • 42min
Building Brain-Like Systems from Sub-Threshold Electronics
Join Professor Giacomo Indiveri, an expert in biologically-plausible neural circuits from the Institute of Neuroinformatics in Zurich. He dives into the complexities of creating brain-like systems using subthreshold analog neuromorphic circuits and discusses the evolution of chip technologies that mimic biological neuron behavior. The conversation highlights how these innovations can enhance consumer devices and improve usability for the elderly, all while navigating the trade-offs between precision and practicality in robotics.
