Close Menu
Technology Mag

    Subscribe to Updates

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

    What's Hot

    The UK’s largest energy supplier has created its own EV charger

    September 3, 2025

    Researchers Are Already Leaving Meta’s New Superintelligence Lab

    September 3, 2025

    Samsung’s new party speakers are less subtle than ever

    September 3, 2025
    Facebook X (Twitter) Instagram
    Subscribe
    Technology Mag
    Facebook X (Twitter) Instagram YouTube
    • Home
    • News
    • Business
    • Games
    • Gear
    • Reviews
    • Science
    • Security
    • Trending
    • Press Release
    Technology Mag
    Home » Selective Forgetting Can Help AI Learn Better
    Business

    Selective Forgetting Can Help AI Learn Better

    News RoomBy News RoomMarch 11, 20243 Mins Read
    Facebook Twitter Pinterest LinkedIn Reddit WhatsApp Email

    The original version of this story appeared in Quanta Magazine.

    A team of computer scientists has created a nimbler, more flexible type of machine learning model. The trick: It must periodically forget what it knows. And while this new approach won’t displace the huge models that undergird the biggest apps, it could reveal more about how these programs understand language.

    The new research marks “a significant advance in the field,” said Jea Kwon, an AI engineer at the Institute for Basic Science in South Korea.

    The AI language engines in use today are mostly powered by artificial neural networks. Each “neuron” in the network is a mathematical function that receives signals from other such neurons, runs some calculations, and sends signals on through multiple layers of neurons. Initially the flow of information is more or less random, but through training, the information flow between neurons improves as the network adapts to the training data. If an AI researcher wants to create a bilingual model, for example, she would train the model with a big pile of text from both languages, which would adjust the connections between neurons in such a way as to relate the text in one language with equivalent words in the other.

    But this training process takes a lot of computing power. If the model doesn’t work very well, or if the user’s needs change later on, it’s hard to adapt it. “Say you have a model that has 100 languages, but imagine that one language you want is not covered,” said Mikel Artetxe, a coauthor of the new research and founder of the AI startup Reka. “You could start over from scratch, but it’s not ideal.”

    Artetxe and his colleagues have tried to circumvent these limitations. A few years ago, Artetxe and others trained a neural network in one language, then erased what it knew about the building blocks of words, called tokens. These are stored in the first layer of the neural network, called the embedding layer. They left all the other layers of the model alone. After erasing the tokens of the first language, they retrained the model on the second language, which filled the embedding layer with new tokens from that language.

    Even though the model contained mismatched information, the retraining worked: The model could learn and process the new language. The researchers surmised that while the embedding layer stored information specific to the words used in the language, the deeper levels of the network stored more abstract information about the concepts behind human languages, which then helped the model learn the second language.

    “We live in the same world. We conceptualize the same things with different words” in different languages, said Yihong Chen, the lead author of the recent paper. “That’s why you have this same high-level reasoning in the model. An apple is something sweet and juicy, instead of just a word.”

    Share. Facebook Twitter Pinterest LinkedIn WhatsApp Reddit Email
    Previous ArticleAirbnb is banning indoor security cameras
    Next Article Russia Attacked Ukraine’s Power Grid at Least 66 Times to ‘Freeze It Into Submission’

    Related Posts

    Researchers Are Already Leaving Meta’s New Superintelligence Lab

    September 3, 2025

    Latam-GPT: The Free, Open Source, and Collaborative AI of Latin America

    September 3, 2025

    Big Tech Companies in the US Have Been Told Not to Apply the Digital Services Act

    September 2, 2025

    Scientists Are Flocking to Bluesky

    September 2, 2025

    Why China Builds Faster Than the Rest of the World

    September 1, 2025

    Anthropic Settles High-Profile AI Copyright Lawsuit Brought by Book Authors

    August 28, 2025
    Our Picks

    Researchers Are Already Leaving Meta’s New Superintelligence Lab

    September 3, 2025

    Samsung’s new party speakers are less subtle than ever

    September 3, 2025

    DOGE Put Everyone’s Social Security Data at Risk, Whistleblower Claims

    September 3, 2025

    Latam-GPT: The Free, Open Source, and Collaborative AI of Latin America

    September 3, 2025
    • Facebook
    • Twitter
    • Pinterest
    • Instagram
    • YouTube
    • Vimeo
    Don't Miss
    News

    Acer’s new Amadana and CE270 monitors are unusually, strikingly stylish

    By News RoomSeptember 3, 2025

    Unfortunately for stateside readers, Acer has no current plans to ship its Amadana branded products…

    Acer’s 16-inch Air weighs even less than a 13-inch MacBook Air

    September 3, 2025

    My brief hands-on with Acer’s new convertible Chromebook has me cautiously optimistic

    September 3, 2025

    Google will reveal its Gemini smart home plans (and hardware) next month

    September 3, 2025
    Facebook X (Twitter) Instagram Pinterest
    • Privacy Policy
    • Terms of use
    • Advertise
    • Contact
    © 2025 Technology Mag. All Rights Reserved.

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