• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI
    Announcement

    ‘Neural Architecture Discovery via Autonomous Evolution’ accepted to COLM 2026

    The author says 1,773 experiments yielded 105 autonomously discovered linear attention architectures with state-of-the-art results under controlled evaluation.

    Weixian XuWX
    1 Source, 5h ago, first seen 5h ago

    TLDR

    The paper’s author announced its acceptance to COLM 2026. They say their system autonomously discovered 105 linear attention architectures across 1,773 experiments and achieved state-of-the-art results under their controlled evaluation. The author also said the project had 1.2K GitHub stars as of October 8, 2026.

    Combined views

    319

    1 Source, first seen 5h ago

    Combined views

    319

    1 Source, first seen 5h ago

    5 likes
    5 likes
    1 comments
    5 saves
    3 reposts
    1 comments
    5 saves
    3 reposts
    Featured Source

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    1 Source

    Weixian Xu@WeixianXuExcited to share that our paper “Neural Architecture Discovery via Autonomous Evolution” has been accepted to COLM 2026! Across 1,773 experiments, our system autonomously discovered 105 linear attention architectures that achieved SOTA results under our controlled evaluation. The project now has 1.2K stars on GitHub and was one of the earliest explorations of autonomous AI for AI: letting AI discover new neural architectures. The field has moved faster than we ever expected, and I’m grateful we could contribute to it! 📍 Poster #56, Imperial Ballroom 🕚 Poster Session 5 · 11:00 AM PDT Unfortunately, I won’t be there in person, but please stop by our poster and feel free to reach out if you’d like to learn more or share feedback. 📄 Paper: https://arxiv.org/abs/2507.18074 💻 Code: https://github.com/GAIR-NLP/ASI-Arch5h