Shikun Liu

@liu_shikun

Ph.D. student at the Dyson Robotics Lab at Imperial College. Study the science of art, and art in science.

London, England
shikun.io
Vrijeme pridruživanja: rujan 2017.
6 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @liu_shikun

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

  1. proslijedio/la je Tweet
    30. sij

    Humans learn from curriculum since birth. We can learn complicated math problems because we have accumulated enough prior knowledge. This could be true for training a ML/RL model as well. Let see how curriculum can help an RL agent learn:

    6 replies 169 proslijeđenih tweetova 810 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. 20. sij

    You will like this one.

    1 reply 1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  3. 20. sij

    Semantic hierarchy exists in deep generative models.

    1 reply 3 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    19. sij

    Some say "cellular automata are just recurrent convolutional networks" (quote , ) so possibly those creatures are already looming inside your self-driving car / Colab notebook

    Somewhere in the hidden layers of your betweeen two convolutions a new life form is just having it's morning coffee. Thanks !
    7 proslijeđenih tweetova 16 korisnika označava da im se sviđa
    Poništi
  5. 19. sij

    A very nice blog discussing the current limitations in deep learning.

    3 proslijeđena tweeta 10 korisnika označava da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    8. sij

    SPACE: Unsupervised Object-Oriented Scene Representation via Spatial Attention and Decomposition “SPACE can explicitly provide factorized object representations for foreground objects while also decomposing background segments of complex morphology.” 👾

    i
    24 proslijeđena tweeta 113 korisnika označava da im se sviđa
    Poništi
  7. proslijedio/la je Tweet
    28. pro 2019.

    Holy crap, this is incredible: A dog playing Jenga. Turn-taking, fine motor control, apparent understanding of the aim of the game. I would not have guessed a dog could do this.

    84.807 proslijeđenih tweetova 295.275 korisnika označava da im se sviđa
    Poništi
  8. proslijedio/la je Tweet
    18. pro 2019.

    [2/2] Find out more about Google’s Project Euphonia and Tim’s story in the new Originals learning series , out today:

    1 reply 60 proslijeđenih tweetova 144 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  9. 19. pro 2019.

    Incredible work.

    View synthesis from a single image by Georgia Gkioxari (FAIR).
    2 korisnika označavaju da im se sviđa
    Poništi
  10. 9. pro 2019.

    Please find me at poster session for a chat on Tuesday afternoon if you are interested in meta learning, auxiliary learning and multi-task learning!

    is presenting our paper at (with Edward Johns) on Self-Supervised Generalisation with Meta Auxiliary Learning. We automatically generate labels for unsupervised auxiliary tasks to improve performance on a primary task.
    2 proslijeđena tweeta 11 korisnika označava da im se sviđa
    Poništi
  12. proslijedio/la je Tweet
    11. stu 2019.

    81 proslijeđeni tweet 258 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet
    10. stu 2019.

    😊Self-supervised learning opens up a huge opportunity for better utilizing unlabelled data while learning in a supervised learning manner. My latest post covers many interesting ideas of self-supervised learning tasks on images, videos & control problems:

    11 replies 487 proslijeđenih tweetova 1.797 korisnika označava da im se sviđa
    Poništi
  14. proslijedio/la je Tweet
    4. stu 2019.

    Adobe Aero has officially arrived. 🤩 Aero is the most intuitive way to blend reality with the surreal in unexpected ways. It's time to create next-gen experiences:

    10 replies 93 proslijeđena tweeta 296 korisnika označava da im se sviđa
    Poništi
  15. 5. stu 2019.

    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  16. 5. stu 2019.

    A very impressive work from Stanford on neural scene representation and rendering.

    1 reply 1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet
    11. lis 2019.

    How should a person exploring exotic geometries be called? A geometronaut?

    301 reply 3.724 proslijeđena tweeta 18.442 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    26. ruj 2019.

    I've open sourced the demos & course notes from my recent "Intro to Creative Coding" workshop on p5.js and Tone.js. 🗒️ Course Notes → 🎨 p5.js Demos → 🔊 Tone.js Demos →

    37 replies 644 proslijeđena tweeta 2.763 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    25. ruj 2019.

    Using gradient descent for everything

    50 replies 519 proslijeđenih tweetova 2.718 korisnika označava da im se sviđa
    Poništi
  20. 25. ruj 2019.

    Just spot an interesting paper at ICLR 2020 which combines adversarial attack and NAS:

    Poništi

@liu_shikun 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:

    ·