Log inSign up
Helen Qu profile banner
user avatar

Helen Qu

@_helenqu
research fellow @FlatironCCA working on ml robustness / multimodal models / AI for science. prev: PhD @physatpenn ‘24, BSE @CIS_Penn '17.
nyc
helenqu.com
Joined May 2012
254
Following
1,202
Followers
RepliesMedia
  • Pinned
    user avatar
    Helen Qu
    @_helenqu
    Dec 23, 2025
    quick update: i started a blog! my first post is live & explains the unstable learning dynamics caused by the "deadly triad" of RL through the lens of basic optimization theory. this was so much fun to learn/write about, hope yall enjoy ✨ helenqu.com/blog/posts/dea… 🧵
    Image
  • user avatar
    Helen Qu
    @_helenqu
    Jul 10
    great to work w @HazanPrinceton on this post!! we contend that the multi-agent context changes what kinds of agents (eg more cooperative agents) are more likely to emerge.
    user avatar
    Elad Hazan
    @HazanPrinceton
    Jul 9
    Bostrom’s orthogonality thesis—the idea that intelligence and goals can be independent—is foundational to alignment theory. Our thoughts on it, in a new essay with @_helenqu (postdoc @ flatiron) and Lauren Li (Princeton student): The Statistical Orthogonality Thesis:
  • user avatar
    Helen Qu
    @_helenqu
    Jul 6
    so hyped to be at icml this week!! 🇰🇷🇰🇷 been thinking a lot about multi-agent systems lately and would love to connect with like-minded folks!
  • user avatar
    Helen Qu
    @_helenqu
    Jun 8
    how should we align multi-agent AI systems? 🤖🤖🤖 inspired by emergent phenomena, i argue for a game theoretic view: can we design incentives that align agents w/ human values, even w/o transparency into agents' intentions? details in my new post! helenqu.com/blog/posts/eme… 🧵
    Image