1. X
  2. Helen Qu ✈️ ICML
Log inSign up
Helen Qu ✈️ ICML
104 posts
Image
user avatar
Helen Qu ✈️ ICML
@_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
253
Following
1,202
Followers
RepliesRepliesMediaMedia

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Pinned
    user avatar
    Helen Qu ✈️ ICML
    @_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
    39K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Apr 17, 2024
    I’m a Dr now!! so grateful to my advisor and all my collaborators, friends, and family for supporting me every step of the way 🥰
    Image
    Image
    94K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Jun 14, 2024
    my thesis is now live on arXiv! arxiv.org/abs/2406.04529
    Image
    Image
    Image
    user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Apr 17, 2024
    I’m a Dr now!! so grateful to my advisor and all my collaborators, friends, and family for supporting me every step of the way 🥰
    64K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Jul 26, 2023
    How can choosing the wrong host galaxy for type Ia supernovae bias cosmology? 🌌 My new paper with the Dark Energy Survey collaboration investigates the impact of host galaxy mismatch on cosmology from 5 years of DES supernova data: arxiv.org/abs/2307.13696, 🧵1/n
    Image
    5.6K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Feb 15, 2024
    Excited to be joining @FlatironCCA @FlatironInst in the fall as a Flatiron Research Fellow! time to say goodbye to philly, my home for basically a decade (!!), and hello to nyc 🏙️
    3.5K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    May 22, 2023
    Can we infer redshift from supernova photometry alone? Introducing Photo-zSNthesis🌱, a deep learning-based approach for type Ia supernova redshift inference. We show >5x improvement on real data over existing SN photo-z approaches! arxiv.org/abs/2305.11869, 🧵 1/5
    Image
    4.8K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Aug 15, 2023
    The PLAsTiCC astronomical time series dataset is now on the Hugging Face Hub! 🤗 Now it only takes one line of code to download and integrate this dataset into your machine learning project - more details in the 🧵: huggingface.co/datasets/helen… 1/5
    Image
    3.8K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Feb 20, 2024
    What's the best way to use unlabeled target data for unsupervised domain adaptation (UDA)? Introducing Connect Later: pretrain on unlabeled data + apply *targeted augmentations* designed for the dist shift during fine-tuning ➡️ SoTA UDA results! arxiv.org/abs/2402.03325 🧵👇
    Image
    12K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Mar 12, 2024
    today, gen AI performance is surprisingly robust to new data/tasks, even beating specialized models! the secret: training on large-scale unlabeled data. what can we as scientists learn from this? some thoughts on robustness & the power of the unlabeled data you already have:
    Image
    2.6K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    May 21, 2023
    I’ll be speaking at the @FlatironCCA @FlatironInst Cosmic Connections workshop on astrophysics x ML this week! Excited to learn about new directions in ML for cosmology, time domain science, and much more 💫
    1.3K
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Oct 20, 2023
    Check out our new overview of transformers/attention and a review of its applications in astronomy! Comments welcome 😊Thanks to @BhuvJain and Dimitrios for all their hard work!
    arXiv logo
    arxiv.org
    Transformers for scientific data: a pedagogical review for astronomers
    The deep learning architecture associated with ChatGPT and related generative AI products is known as transformers. Initially applied to Natural Language Processing, transformers and the...
    387
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Aug 15, 2023
    Replying to @_helenqu
    Finally, all credit goes to the original developers of PLAsTiCC incl @reneehlozek @lgalbany @emilleishida @gsnarayan @_sublunar_ and many others! I hope this is a step towards making ML in astro more accessible and reproducible. Feel free to reach out if you run into issues! 5/5
    289
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    May 22, 2023
    Replying to @_helenqu
    We show a >5x improvement on mean residuals over the widely used SN photo-z predictor, LCFIT+Z, on simulated/real SDSS + simulated LSST data! We also find that performance on real data is slightly diminished (compared to simulations) but still much better than the baseline. 4/5
    Image
    375
  • user avatar
    Helen Qu ✈️ ICML
    @_helenqu
    Mar 12, 2024
    Replying to @_helenqu
    Is pretraining all you need? We found that it sometimes fails to boost performance beyond no pretraining… We figured out why and developed Connect Later to boost accuracy and robustness with pretraining in all cases! arxiv.org/abs/2402.03325 9/10
    arXiv logo
    arxiv.org
    Connect Later: Improving Fine-tuning for Robustness with Targeted...
    Models trained on a labeled source domain (e.g., labeled images from wildlife camera traps) often generalize poorly when deployed on an out-of-distribution (OOD) target domain (e.g., images from...
    666
Advertisement
Advertisement