Jacob Menick

@jacobmenick

Deep Learning researcher at . I speak only for myself.

London
Vrijeme pridruživanja: kolovoz 2018.
11 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @jacobmenick

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @jacobmenick

  1. proslijedio/la je Tweet
    31. sij

    When led an effort to change the debate thresholds, the DNC refused—saying they couldn’t benefit any candidate. It seems the only candidate they’re willing to benefit is a billionaire who’s buying his way into the race. Total mess.

    The DNC drastically shifts debate requirements, opening the door for Bloomberg to join the stage starting in mid-February
    1.838 replies 13.369 proslijeđenih tweetova 55.931 korisnik označava da mu se sviđa
    Poništi
  2. proslijedio/la je Tweet
    29. sij

    A robot in Germany shows that machines can learn to do the job of a human (*learn* being the key word): (with the great )

    1 reply 26 proslijeđenih tweetova 72 korisnika označavaju da im se sviđa
    Poništi
  3. 21. sij

    Machine learning is not about finding the best model for a given task/dataset. It's about jointly optimizing over model/task/dataset for the desired capabilities of a computer program.

    1 reply 1 proslijeđeni tweet 26 korisnika označava da im se sviđa
    Poništi
  4. 19. sij

    Apache beam reax only

    1 reply 3 korisnika označavaju da im se sviđa
    Poništi
  5. 3. sij

    Awesome work and interesting thread by Jeffrey.

    Very excited to share where we show an AI system that outperforms specialists at detecting breast cancer during screening in both the UK and US. Joint work with and published in today!
    Prikaži ovu nit
    3 korisnika označavaju da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    20. pro 2019.

    Really happy and a lovely early-Christmas present that our paper "Multiplicative Interactions and Where to Find Them" was accepted to . We analyse such interactions, unify different forms (eg hypernets, gating), and encourage you to use more of them! 1/2

    8 proslijeđenih tweetova 76 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  7. 27. stu 2019.

    related work

    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. 27. stu 2019.

    New work by Utku Evci et al. on sparse training. My contribution was helping with the RNN experiments. Fun collaborating with and getting involved in sparse man 's sweeping sparsity research programme.

    80% sparse ResNet-50
    End-to-end training of sparse deep neural networks with little-to-no performance loss. Check out our new paper: “Rigging the Lottery: Making All Tickets Winners” (RigL👇) ! 📃 📁 with and
    Prikaži ovu nit
    2 proslijeđena tweeta 12 korisnika označava da im se sviđa
    Poništi
  9. 27. stu 2019.

    accompanying 45

    Everyone is a winner 🔥
    1 reply 1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    6. stu 2019.

    Not sure, who made this, but thanks! 🤣🤣🤣

    19 replies 302 proslijeđena tweeta 1.447 korisnika označava da im se sviđa
    Poništi
  11. 14. lis 2019.

    A core aspect of deep learning research is removing degrees of freedom. We only know what works from running experiments but the architecture search space is combinatorial, so the only practical way to innovate is to remove degrees of freedom and tinker within a confined space.

    1 proslijeđeni tweet 35 korisnika označava da im se sviđa
    Poništi
  12. 17. ruj 2019.

    One thing that unifies all humanity: every person thinks they listen to sick music

    1 reply 2 korisnika označavaju da im se sviđa
    Poništi
  13. 6. ruj 2019.

    Dr. Pfau strikes back!

    Thrilled to be able to share what I've been working on for the last year - solving the fundamental equations of quantum mechanics with deep learning!
    Prikaži ovu nit
    3 korisnika označavaju da im se sviđa
    Poništi
  14. 8. kol 2019.

    RIP houghton 2019

    1 reply
    Poništi
  15. proslijedio/la je Tweet
    15. srp 2019.

    DVD-GAN

    0:01
    Efficient Video Generation on Complex Datasets pdf: abs:
    Prikaži ovu nit
    7 replies 98 proslijeđenih tweetova 382 korisnika označavaju da im se sviđa
    Poništi
  16. 12. srp 2019.

    Kuuul

    We scaled up neural ODE time series models to real data, such as medical records with irregularly-timed measurements. Specifically, we made ODE-RNN hybrids, and also improved inference in latent ODEs: By amazing students and .
    Prikaži ovu nit
    Poništi
  17. 3. srp 2019.

    Stand back I'm about to do some wild shit using a technique called machine learning

    1 reply 8 korisnika označava da im se sviđa
    Poništi
  18. 22. lip 2019.

    Don’t model the data, model the dataset :-)

    Excited to share our latest work; it combines neural networks with stochastic processes by positing priors over the relational structure of the dataset, without relying on global latent variables!
    4 korisnika označavaju da im se sviđa
    Poništi
  19. 6. lip 2019.

    Nice progress in generative models by colleagues

    VQVAE-2 finally out! Powerful autoregressive models in a hierarchical compressed latent space. No modes were collapsed in the creation of these samples ;) Arixv: With and More samples and details 👇 [thread]
    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    31. svi 2019.

    Medical images can exhibit ambiguities on multiple scales & locations often varying independently. We propose a hierarchical generative model to capture such variations in segmentations & show much improved sample fidelity and fit with the GT distribution:

    33 proslijeđena tweeta 110 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi

@jacobmenick još nije objavio nijedan Tweet.

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·