Log inSign up
Mitchell Ostrow
164 posts
Image
user avatar
Mitchell Ostrow
@neurostrow
PhD Candidate in Computational Neuroscience + ML at @mitbrainandcog, prev @yaleneuro, @yaledatascience, @meta. NSF GRFP Fellow
mitchellostrow.github.io
Joined September 2022
527
Following
1,385
Followers
  • Pinned
    user avatar
    Mitchell Ostrow
    @neurostrow
    Nov 28, 2025
    We presented DSA at NeurIPS in 2023 (more extensions to come, stay tuned!) x.com/neurostrow/sta…
    user avatar
    Mitchell Ostrow
    @neurostrow
    Nov 10, 2025
    Replying to @neurostrow
    The basic idea of DSA: approximate your dynamics so that comparison is tractable. This is backed by Koopman Operator Theory and relates to work done by @wredman4 , @IgorMezic InputDSA naturally extends DSA—we can compare intrinsic dynamics, effect of input, or both jointly! 3/
    Image
    1.9K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Nov 22, 2024
    Applying to grad school is hard. The expectations for essays are murky, and not everybody comes from a wealthy institution with resources needed to optimize an application. If you're applying to grad school now, please reach out and I'll be happy to help you edit your essays.
    124K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 21, 2023
    I’m excited to share the first paper of my PhD! arxiv.org/abs/2306.10168 1/
    Image
    72K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 23, 2024
    How can we understand the way sequence models choose to combine info in their context to make good predictions? In our new @icmlconf workshop paper, @adamjeisen , @FieteGroup and I provide a new theoretical lens from dynamical systems: delay embeddings arxiv.org/abs/2406.11993
    Image
    13K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Sep 22, 2023
    Accepted to @NeurIPSConf!! Looking forward to presenting in New Orleans this December 🧠💻
    user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 21, 2023
    I’m excited to share the first paper of my PhD! arxiv.org/abs/2306.10168 1/
    Image
    18K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Dec 12, 2023
    I'll be presenting this work today (Tuesday) in the 5:15-7:15pm poster session, #416! Come through if you want to chat about Koopman operators, neural manifolds / dynamics, or just to say hi!
    user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 21, 2023
    I’m excited to share the first paper of my PhD! arxiv.org/abs/2306.10168 1/
    Image
    6.4K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Nov 23, 2024
    Wow, this blew up! I'll try to get to everyone's requests but please be patient.
    user avatar
    Mitchell Ostrow
    @neurostrow
    Nov 22, 2024
    Applying to grad school is hard. The expectations for essays are murky, and not everybody comes from a wealthy institution with resources needed to optimize an application. If you're applying to grad school now, please reach out and I'll be happy to help you edit your essays.
    2.4K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Nov 23, 2024
    Replying to @neurostrow
    Dms are unlocked if you want to contact me there, but please be patient.
    4.1K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Aug 25, 2023
    I'll be presenting this work @CogCompNeuro as a poster (P-3B:107) and as a contributed talk (CT-2). Come say hi!
    user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 21, 2023
    I’m excited to share the first paper of my PhD! arxiv.org/abs/2306.10168 1/
    Image
    1.5K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 21, 2023
    Replying to @neurostrow
    DSA combines recent advances in data-driven dynamical systems theory and shape analysis to tease apart dynamics and geometry. We learn a high-dimensional linear embedding of nonlinear systems using HAVOK, a variant of Dynamic Mode Decomposition from @eigensteve. 3/
    Image
    1.7K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 21, 2023
    Replying to @neurostrow
    We don’t view DSA as a replacement for current similarity metrics but rather as a complement that emphasizes a distinct component of neural computation—time. 9/
    809
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Jun 21, 2023
    Replying to @neurostrow
    A paper from @ItsNeuronal and @niru_m showed that different RNN architectures trained on this task have the same attractor topology, but statistical shape metrics can distinguish them. We replicate their result and show that DSA identifies each architecture as equivalent. 5/
    Image
    1K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Aug 16, 2024
    Had a lot of fun writing about @courellis 's and @UeliRutishauser 's amazing research!
    user avatar
    Mary Elizabeth
    @meharpist
    Aug 16, 2024
    Replying to @meharpist
    It's accompanied by an excellent @NatureNV by @neurostrow and Ila Fiete nature.com/articles/d4158…
    2K
  • user avatar
    Mitchell Ostrow
    @neurostrow
    Nov 23, 2024
    Replying to @neurostrow
    Very appreciative of all the interest surrounding this post! I’ll do my best to get through them all but for now I have to ask y’all to hold off contacting me for a bit.
    2K

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
Advertisement
Advertisement