George Papamakarios

@gpapamak

Research scientist at DeepMind

London, England
gpapamak.github.io
Vrijeme pridruživanja: prosinac 2017.
5 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @gpapamak

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

  1. Prikvačeni tweet
    6. pro 2019.

    Check out our extensive review paper on normalizing flows! This paper is the product of years of thinking about flows: it contains everything we know about them, and many new insights. With , , , . Thread 👇

    Looking for something to read in your flight to ? Read about Normalizing Flows from our extensive review paper (also with new insights on how to think about and derive new flows) with
    Prikaži ovu nit
    174 proslijeđena tweeta 630 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    25. pro 2019.

    Some outcomes: Accepted (spotlight): Hamiltonian Generative Networks, Rejected: Causally Correct Partial Models for Reinforcement Learning Congrats to all my collaborators on both, independently of acceptance!

    27 proslijeđenih tweetova 163 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    9. pro 2019.

    I implemented some normalizing flows yesterday (NICE, RealNVP, MAF, IAF), tried to make core of it somewhat clean in case helpful I like how flow layers can be structured similar to backprop, each needs an invert() and emits a log det J "regularization"

    14 replies 133 proslijeđena tweeta 761 korisnik označava da mu se sviđa
    Poništi
  4. proslijedio/la je Tweet
    8. pro 2019.

    Find myself, , and at poster #117 on Wednesday 10:45 - 12:45 in Hall B + C to talk flows. Disclaimer: You will be quizzed mercilessly on the survey paper you were meant to read on the plane over.

    12 proslijeđenih tweetova 44 korisnika označavaju da im se sviđa
    Poništi
  5. proslijedio/la je Tweet
    8. pro 2019.

    we'll have open-mic sessions to ignite discussions, as well as invited spotlights! First one will be ( & ) talking about prospects and challenges of Inference with check his survey with

    1 reply 3 proslijeđena tweeta 12 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  6. proslijedio/la je Tweet
    6. pro 2019.

    Normalizing Flows let you build up complex, yet still easy to work with probability distributions. Want to learn more? Check out this video I made covering the basics of this growing class of techniques with an example application in generative modeling.

    5 replies 74 proslijeđena tweeta 316 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  7. 6. pro 2019.

    We hope there is something there for everyone interested in flows: - A gentle introduction for those wanting to get started. - Explanations of existing flows for practitioners who want to deepen their understanding. - Advanced topics for seasoned experts.

    3 proslijeđena tweeta 31 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    21. stu 2019.

    This feels like a real breakthrough: Take the same basic algorithm as AlphaZero, but now *learning* its own simulator. Beautiful, elegant approach to model-based RL. ... AND ALSO STATE OF THE ART RESULTS! Well done to the team at

    5 replies 209 proslijeđenih tweetova 802 korisnika označavaju da im se sviđa
    Poništi
  9. proslijedio/la je Tweet
    12. stu 2019.

    New: "Training deep neural density estimators to identify mechanistic models of neural dynamics” Nonnenmacher Bassetto . Our biggest project so far! Thread:

    82 proslijeđena tweeta 178 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    6. stu 2019.

    In our newest paper we discuss the frontier of simulation-based inference (aka likelihood-free inference) for a broad audience. We identify three main forces driving the frontier including: , active learning, and integration of autodiff and probprog.

    6 replies 52 proslijeđena tweeta 164 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    4. stu 2019.

    Interested in flows for ordinal discrete data? We have released the code for our paper on Integer Discrete Flows and Lossless Compression. Check it out at . In collaboration with , , and .

    Check out our work on Integer Discrete Flows, generative flows for ordinal discrete data. In short: no (de)quantization, no Jacobian computation, and Lossless Compression. (link: ). In collaboration with , , and .
    38 proslijeđenih tweetova 131 korisnik označava da mu se sviđa
    Poništi
  12. 30. lis 2019.

    It contains, among other things: - A tutorial on density estimation. - A tutorial on approximate Bayesian computation (ABC). - A review of normalizing flows (up until April 2019). - Extensive commentary on the papers I published as part of my PhD.

    1 reply 2 proslijeđena tweeta 39 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  13. 30. lis 2019.

    My PhD thesis is now available on arXiv: Neural Density Estimation and Likelihood-free Inference There's a lot in it for those interested in probabilistic modelling with normalizing flows, and in likelihood-free inference using machine learning. (cont.)

    Congratulations to for passing his PhD viva! George's dissertation is a great example of how to include published papers and put them in context. Recommended reading for density estimation with neural networks, and likelihood-free inference.
    9 replies 144 proslijeđena tweeta 776 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    4. ruj 2019.

    Don't forget you can satisfy many of your flow-based modeling needs with this handy PyTorch library we released alongside the paper .

    11 proslijeđenih tweetova 31 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    4. ruj 2019.

    Neural Spline Flows accepted to ! Come and talk to us in Vancouver. Work with , , . Paper: . Code: (includes a nice framework for building flows).

    Happy to announce new work with three amazing co-authors. We show that flow-based density models with coupling transforms can be made surprisingly flexible by using monotonic splines and invertible linear layers. Experiments include density estimation, VAEs and images.
    1 reply 23 proslijeđena tweeta 105 korisnika označava da im se sviđa
    Poništi
  16. proslijedio/la je Tweet
    27. srp 2019.

    I had a great time examining this thesis and learned a lot. In addition to the published papers, the thesis has a brilliant overview of the normalising flows literature (up to a few months ago!)

    Congratulations to for passing his PhD viva! George's dissertation is a great example of how to include published papers and put them in context. Recommended reading for density estimation with neural networks, and likelihood-free inference.
    2 proslijeđena tweeta 21 korisnik označava da mu se sviđa
    Poništi
  17. proslijedio/la je Tweet
    27. srp 2019.

    Congratulations to for passing his PhD viva! George's dissertation is a great example of how to include published papers and put them in context. Recommended reading for density estimation with neural networks, and likelihood-free inference.

    33 proslijeđena tweeta 275 korisnika označava da im se sviđa
    Poništi
  18. proslijedio/la je Tweet
    15. lip 2019.

    Slides & Code for my Normalizing Flow tutorial at ICML here:

    56 proslijeđenih tweetova 265 korisnika označava da im se sviđa
    Poništi
  19. proslijedio/la je Tweet
    14. lip 2019.

    I'll be speaking about Neural Spline Flows from 12:10-12:30 tomorrow at the Invertible Neural Networks and Normalizing Flows workshop at . Myself and will be around all day, so drop by our poster for a chat!

    1 reply 3 proslijeđena tweeta 21 korisnik označava da mu se sviđa
    Poništi
  20. proslijedio/la je Tweet
    13. lip 2019.

    Important announcements for attendees of our workshop on Invertible neural networks and normalizing flows (INNF) **this Saturday at ICML2019 (room 103)**: 1) BRING YOUR LAPTOP for the tutorial! (9:30am) 2) Submit questions for the panel! (5pm):

    12 proslijeđenih tweetova 26 korisnika označava da im se sviđa
    Poništi

@gpapamak 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:

    ·