Kai Arulkumaran

@kaixhin

Researcher, programmer, DJ, transhumanist. Currently ; formerly research intern ///, mentor .

London, UK
Vrijeme pridruživanja: kolovoz 2011.

Tweetovi

Blokirali ste korisnika/cu @kaixhin

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

  1. Prikvačeni tweet
    21. kol 2017.

    A Brief Survey of Deep Reinforcement Learning w/ & :

    Poništi
  2. proslijedio/la je Tweet
    prije 20 sati

    François () is both an independent ML researcher investigating sparse efficient models and... an early angel investor in 🤗 Happy that he agreed to share some of his knowledge and experience on sparse models in a series of posts! First one is here

    Poništi
  3. 4. velj

    Just tried to do "From Zero to MuZero" in 30 minutes (zero = starting from RL) for my lab meeting! Ran through MDPs, RL loop, MB vs. MF RL, Bellman backups, MCTS, POMDPs, and then how MuZero combines these to learn a latent state to run AlphaZero style.

    Poništi
  4. 2. velj
    Poništi
  5. proslijedio/la je Tweet
    31. sij

    On Tuesday, in my class, we have learnt that all a neural net does is stretching / contracting the space fabric. For example this 3-layer net (1 hidden layer of 100 positive neurons) gets its 5D logits (2D projections) linearly separable by the classifier hyperplanes (lines).

    Poništi
  6. proslijedio/la je Tweet
    1. velj

    This is a great opportunity to get involved and support . Apply to be an organizer or share with anyone who could be interested!

    Poništi
  7. 31. sij

    Excited to see what has achieved, and there's more details about their approach here. Seems that scaling up and combining current methods (SL/IL/RL) in smart ways can go pretty far!

    Poništi
  8. proslijedio/la je Tweet
    30. sij

    We're standardizing OpenAI's deep learning framework on PyTorch to increase our research productivity at scale on GPUs (and have just released a PyTorch version of Spinning Up in Deep RL):

    Poništi
  9. 30. sij

    Fantastic initiative from combining a range of AI techniques to help blind/partially sighted people! Happy to have played a small role in the past.

    Poništi
  10. proslijedio/la je Tweet
    29. sij

    After 2+ years in stealth, we’re excited to launch today! Thank you to our team, customers, partners and investors, we couldn’t have done it w/o your support and trust. Exciting milestone, even more exciting journey ahead!

    Poništi
  11. proslijedio/la je Tweet
    29. sij

    As someone who didn't go for it, this is a great view into the life of a phd student from 👀. Love this in particular 👉 Everyone has a different “work-life balance” (and for me, “work” is an essential part of my “life”)

    Poništi
  12. 29. sij

    I did an interview with about my work, my life, and my advice for PhDs. So please check it out if you'd like to find out about any of those 🙂

    Poništi
  13. 24. sij

    Get you a model class that can approximate it all 😉

    Poništi
  14. 23. sij

    Is there a word for the feeling you get after watching a Makoto Shinkai film? Just saw Weathering With You, and feeling like that.

    Poništi
  15. 23. sij

    I was hoping to ask him what the reward function for AGI was, but someone basically beat me to it. David's response to "where do rewards come from?" was that he was concerned with solving specified problems, and left that question to other people.

    Prikaži ovu nit
    Poništi
  16. 22. sij

    David Silver's "provocative" slide. A "conjecture" that this might take us towards the goal of AGI (but not the "hubris" that this is definitely the way to go).

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet
    22. sij

    You got a bunch of GPU machines and are wondering which GPUs are still free? I used to ssh into the list of machines and checked manually until I built a small dashboard to show me: I thought other people might have the same problem..

    Poništi
  18. proslijedio/la je Tweet
    15. sij

    Differentiable Digital Signal Processing (DDSP)! Fusing classic interpretable DSP with neural networks. ⌨️ Blog: 🎵 Examples: ⏯ Colab: 💻 Code: 📝 Paper: 1/

    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    15. sij

    v1.4: customizable mobile builds, Distributed Model Parallelism via experimental RPC API, Java Bindings, Chaining LRSchedulers Summary: Release Notes: Last release for Python 2 (bye bye!)

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    13. sij

    Introducing computer-designed organisms. Our new study out this week in PNAS. w/ , Douglas Blackiston, and

    Prikaži ovu nit
    Poništi
  21. proslijedio/la je Tweet
    14. sij

    I often meet research scientists interested in open-sourcing their code/research and asking for advice. Here is a thread for you. First: why should you open-source models along with your paper? Because science is a virtuous circle of knowledge sharing not a zero-sum competition

    Prikaži ovu nit
    Poništi

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:

    ·