Log inSign up
Eero Simoncelli
164 posts
user avatar
Eero Simoncelli
@EeroSimoncelli
Professor at NYU; Scientific Director, Ctr for Computational Neurocience, Flatiron Institute. Research in Computational Vision (neurons, perception, machines).
New York, NY
cns.nyu.edu/~eero
Joined February 2022
162
Following
1,889
Followers

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
    Eero Simoncelli
    @EeroSimoncelli
    Feb 25
    Replying to @docmilanfar
    Interesting article! In complementary work we recently proved that the self-scheduled diffusion method we developed in july 2020 [arXiv:2007.13640] uses a blind denoiser to correctly sample from the learned data distribution: arXiv:2602.09639
    1.8K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Oct 25, 2025
    Looking for a PhD program where you can study vision from different perspectives (computer vision, neuroscience, perception, image processing)? Consider applying to NYU! visionscience.com/pipermail/visi…
    36K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Oct 7, 2024
    Looking for a PhD program where you can study vision from different perspectives (computer vision, neuroscience, perception, image processing)? Consider applying to NYU! visionscience.com/pipermail/visi…
    69K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Jul 10, 2024
    These two remarkable formulas, derived for the case of additive Gaussian noise, can be generalized to many other of noise: cns.nyu.edu/~lcv/pubs/make…
    user avatar
    Gabriel Peyré
    @gabrielpeyre
    Jul 8, 2024
    Replying to @gabrielpeyre
    Using the SURE, it is straightforward to derive Tweedie’s formula for the optimal denoiser as the score function (the gradient of the log-likelihood of the noisy data), which is used in diffusion models.
    30K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Jan 14, 2025
    Graduate students and advanced undergraduates: Interested in a 3-month summer research internship in Computational Neuroscience @FlatironCCN Come join us in June! Application deadline 17 Jan 2025: apply.interfolio.com/159680
    17K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Jan 20, 2024
    Summer Research Internships in Computational Neuroscience @FlatironCCN Graduate students and advanced undergraduates: Spend a summer doing research at the Flatiron Institute in NYC! Application deadline 2 Feb 2024: apply.interfolio.com/137388
    25K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Nov 24, 2024
    The Center for Neural Science at NYU is aiming to add another computational neuroscientist to our community: apply.interfolio.com/157767 Come join us!
    38K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Oct 16, 2025
    Semantic categories are captured in a union-of-subspaces deep within an unsupervised unconditional denoiser!
    user avatar
    Zahra Kadkhodaie
    @ZKadkhodaie
    Oct 16, 2025
    Diffusion models learn probability densities by estimating the score with a neural network trained to denoise. What kind of representation arises within these networks, and how does this relate to the learned density? @EeroSimoncelli @StephaneMallat and I explored this question.
    19K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Oct 8, 2024
    Looking for a PhD program where you can study computational neuroscience? NYU has a fantastic array of researchers covering the field: groups.google.com/g/systems-neur…
    9.1K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Jun 9, 2025
    Our latest work on the longstanding problem of learning a (normalized) energy model from data…
    user avatar
    Florentin Guth
    @FlorentinGuth
    Jun 6, 2025
    What is the probability of an image? What do the highest and lowest probability images look like? Do natural images lie on a low-dimensional manifold? In a new preprint with @ZKadkhodaie @EeroSimoncelli, we develop a novel energy-based model in order to answer these questions: 🧵
    7.8K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Jul 23, 2024
    Surprising and fun that this abstract object underlies real-world data! A more intuitive version: the 2D manifold of local orientation and phase for windowed sinusoids (“Gabor functions”) is a Klein bottle:
    user avatar
    Peyman Milanfar
    @docmilanfar
    Jul 21, 2024
    Images aren’t arbitrary collections of pixels -they have complicated structure, even small ones. That’s why it’s hard to generate images well. Let me give you an idea: 3×3 gray images represented as points in ℝ⁹ lie approximately on a 2-D manifold: the Klein bottle! 1/3
    11K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Jun 12, 2024
    Just out, from @Jingyang_zhou and @lyndoryndo at @FlatironCCN and @NYU_CNS Co-existence of Weber’s Law, Steven’s Law, Fechner’s Law (and their generalizations). DOI: 10.1073/pnas.2312293121
    user avatar
    Jingyang Zhou
    @Jingyang_zhou
    Jun 10, 2024
    Replying to @Jingyang_zhou
    @FlatironInst @FlatironCCN
    8.6K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Dec 7, 2023
    Congratulations to our @FlatironCCN members who will be presenting work at @NeurIPSconf next week - Looking forward to {sharing, discussing, learning} in New Orleans!
    7.2K
  • user avatar
    Eero Simoncelli
    @EeroSimoncelli
    Aug 27, 2025
    Determining optimal partial linear measurements for a signal class is a foundational problem in signal processing. For Gaussian-distributed signals, the solution is PCA. We compute (empirically) a solution for the signal density learned by a denoiser (diffusion model):
    3.8K