Close Menu
    Main Menu
    • Home
    • News
    • Tech
    • Robotics
    • ML & Research
    • AI
    • Digital Transformation
    • AI Ethics & Regulation
    • Thought Leadership in AI

    Subscribe to Updates

    Get the latest creative news from FooBar about art, design and business.

    What's Hot

    Video games for Change provides 5 new leaders to its board

    June 9, 2025

    Constructing clever AI voice brokers with Pipecat and Amazon Bedrock – Half 1

    June 9, 2025

    ChatGPT’s Reminiscence Restrict Is Irritating — The Mind Reveals a Higher Method

    June 9, 2025
    Facebook X (Twitter) Instagram
    UK Tech Insider
    Facebook X (Twitter) Instagram Pinterest Vimeo
    UK Tech Insider
    Home»News»A memristor-based Bayesian machine
    News

    A memristor-based Bayesian machine

    Amelia Harper JonesBy Amelia Harper JonesMay 13, 2025No Comments4 Mins Read
    Facebook Twitter Pinterest Telegram LinkedIn Tumblr Email Reddit
    A memristor-based Bayesian machine
    Share
    Facebook Twitter LinkedIn Pinterest Email Copy Link


    Over the previous few a long time, the efficiency of machine studying fashions on varied real-world issues has improved considerably. Nevertheless, coaching these fashions, in addition to their implementation, nonetheless requires an enormous quantity of power and computing energy.

    Subsequently, engineers world wide are attempting to develop various {hardware} options that may run synthetic intelligence fashions extra effectively. This may occasionally contribute to their widespread use and enhance their sustainability. A few of these options are based mostly on memristors, reminiscence gadgets that may retailer data with out consuming energy.

    Memristors are digital elements based mostly on nanodevices that restrict or regulate the circulate of electrical present in a circuit, and in addition register upfront how a lot power has handed via it. As a result of they carry out each computation and knowledge storage, these gadgets can higher carry out data processing duties just like the human mind.

    Researchers at Université Paris-Saclay- CNRS, Université Grenoble-Alpes-CEA-LETI, HawAI.tech, Sorbonne Université, and Aix-Marseille Université-CNRS have developed a prototype Bayesian machine (i.e. a tool based mostly on Bayes’ theorem) utilizing memristors. The system they proposed, that was introduced in a paper printed in Nature Electronics, turned out to be considerably extra power environment friendly than presently used {hardware} options.

    Picture of an entire Bayesian machine below an optical microscope

    Based on Damien Querlioz, one of many researchers who carried out the research: “Synthetic intelligence is making main progress at the moment however faces a problem: its appreciable power consumption. It’s now properly understood that this consumption comes from the separation, in computer systems, between computation and reminiscence capabilities. As synthetic intelligence makes use of loads of knowledge, it requires loads of reminiscence, which is expensive to entry when it comes to power. Our brains are far more power environment friendly as a result of the reminiscence capabilities are built-in as shut as doable to the computation capabilities, and we wished to breed this technique”.

    “Till lately, memristors had been an rising expertise, and we couldn’t understand full methods with them,” – Querlioz defined. “Now, our staff constructed a Bayesian machine, a small synthetic intelligence with memristors. The prototype includes 2048 hafnium oxide memristors and 30080 silicon transistors (MOSFETs)”.

    A zoomed-in optical microscopic picture of the Bayesian machine on considered one of its 16 memristor arrays. Picture: Damien Querlioz (CNRS/College Paris-Saclay)

    The structure of the machine is designed by writing Bayes’ regulation in such a means that its implementation is pure in accordance with the ideas of distributed reminiscence and stochastic computing, permitting the circuit to operate completely utilizing native reminiscence and minimal knowledge motion. The researchers use a hybrid complementary metal-oxide-semiconductor/memristor course of and show that the bigger machine design is extra power environment friendly within the sensible activity of gesture recognition than the usual implementation of Bayesian inference on a microcontroller. It might probably acknowledge sure human gestures utilizing hundreds of occasions much less energy than a standard {hardware} resolution based mostly on a microcontroller. Additionally, this Bayesian machine supplies prompt on/off and is immune to single failures.

    “A lot of the analysis on memristor-based machine studying goals at implementing deep studying,” – Querlioz mentioned. “That is, in fact, a particularly necessary aim, as deep studying is so profitable at the moment. Nevertheless, deep studying has some limitations: its outcomes are usually not explainable, and it doesn’t carry out properly when little knowledge is on the market. Right here, we selected to implement Bayesian reasoning, another AI strategy that doesn’t do properly in huge knowledge purposes the place deep studying works so properly, however excels in small knowledge conditions, and supplies absolutely explainable outcomes”.

    Sooner or later, the memristor-based Bayesian machine created by this group of researchers might assist enhance the ability effectivity of AI fashions, in addition to doubtlessly encourage the event of different related options. This may be notably helpful for safety-critical purposes equivalent to medical sensors and industrial security monitoring circuits.

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    Amelia Harper Jones
    • Website

    Related Posts

    ChatGPT’s Reminiscence Restrict Is Irritating — The Mind Reveals a Higher Method

    June 9, 2025

    Stopping AI from Spinning Tales: A Information to Stopping Hallucinations

    June 9, 2025

    Why Gen Z Is Embracing Unfiltered Digital Lovers

    June 9, 2025
    Top Posts

    Video games for Change provides 5 new leaders to its board

    June 9, 2025

    How AI is Redrawing the World’s Electrical energy Maps: Insights from the IEA Report

    April 18, 2025

    Evaluating the Finest AI Video Mills for Social Media

    April 18, 2025

    Utilizing AI To Repair The Innovation Drawback: The Three Step Resolution

    April 18, 2025
    Don't Miss

    Video games for Change provides 5 new leaders to its board

    By Sophia Ahmed WilsonJune 9, 2025

    Video games for Change, the nonprofit group that marshals video games and immersive media for…

    Constructing clever AI voice brokers with Pipecat and Amazon Bedrock – Half 1

    June 9, 2025

    ChatGPT’s Reminiscence Restrict Is Irritating — The Mind Reveals a Higher Method

    June 9, 2025

    Stopping AI from Spinning Tales: A Information to Stopping Hallucinations

    June 9, 2025
    Stay In Touch
    • Facebook
    • Twitter
    • Pinterest
    • Instagram
    • YouTube
    • Vimeo

    Subscribe to Updates

    Get the latest creative news from SmartMag about art & design.

    UK Tech Insider
    Facebook X (Twitter) Instagram Pinterest
    • About Us
    • Contact Us
    • Privacy Policy
    • Terms Of Service
    • Our Authors
    © 2025 UK Tech Insider. All rights reserved by UK Tech Insider.

    Type above and press Enter to search. Press Esc to cancel.