Hardware Podcasts

How Memristors Will Help Machines Think at Different Timescales

EE Times Current · 1 Dec 2023 · 50m

What this episode covers

Professor Melika Payvand of the Institute of Neuroinformatics in Zurich explains how memristors enable neuromorphic systems to operate across multiple time scales, addressing the von Neumann bottleneck by storing neuron parameters locally. She details using resistive memory to implement dendritic delays and structural plasticity, which reduces power consumption by two orders of magnitude compared to recurrent networks. The discussion with hosts Dr. Sunny Bains and Dr. Giulia D’Angelo, and commentator Professor Ralph Etienne-Cummings, covers the evolution from DARPA’s MURI project to current commercial memristor applications by IBM and Micron.

Most relevant to

early-stage companies

Topics

Mentioned

Companies
Institute of Neuroinformatics, DARPA, IBM, Micron, STMicroelectronics
Products
TrueNorth, Loihi, Pimera, SpinR

Related episodes from other shows

Other shows on the same ground, closest first.

  • Moore's Lobby · 49m

    Marveling at MEMS: The New Superheroes of The Silicon World

    Fari Assaderaghi, EVP of Technology and Engineering at SiTime, discusses the evolution of MEMS resonators from low-end timing references to high-performance components that outperform quartz crystals. He details SiTime's strategy of co-developing specialized silicon processes with manufacturing partners to achieve 100,000x improvements in stability and phase noise. Assaderaghi also reflects on his career path through IBM and NXP, emphasizing the importance of first-principles thinking and focusing on ultimate goals rather than perceived limitations.

    Chips, FPGAs & semiconductorsManufacturing partners & offshoring
    7 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by EE Times Current. The episode itself belongs to its publisher.