Scott Jordan

@UMassScott

PhD candidate at UMASS studying reinforcement learning and representation learning

Vrijeme pridruživanja: listopad 2016.

Tweetovi

Blokirali ste korisnika/cu @UMassScott

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

  1. 30. sij

    I agree completely. The practice of making claims, particularly in regards to performance, without sufficient support needs to end.

    Poništi
  2. proslijedio/la je Tweet

    I have a favor to ask. For 20+ years I've been working on a dream: to make all science funded by US taxpayers freely available to all. We are on the verge of achieving this. But we need to show that people care. So please, if you can, sign this letter:

    Poništi
  3. proslijedio/la je Tweet
    21. stu 2019.

    Delighted to share our Science article on making it easier ensure AI systems satisfy societal values. Lead by former postdoc Phil Thomas, w/Castro da Silvam, Barto, Giguere, Brun.

    Poništi
  4. proslijedio/la je Tweet
    21. stu 2019.

    In research published today in , researchers Philip Thomas, & Andy Barto + colleagues from , introduce a new “Seldonian” framework for fairer, safer algorithms:

    Poništi
  5. proslijedio/la je Tweet
    23. lis 2019.

    Pretty good article discussing 's Rubik's Cube result: This is what good criticism looks like, unlike most of the critical tweets mentioned in this article. Good work and .

    Poništi
  6. proslijedio/la je Tweet

    Philip Thomas is the Asst Professor and co-director of the Autonomous Learning Lab at . He shared some of his upcoming work on a new framework for designing ML algorithms. View slides from his "Safe and Fair Reinforcement Learning" talk here:

    Philip Thomas  presents his work to a large seated audience at RL Day 2019. Attendees are watching the screen behind him.
    Philip Thomas points at the projected screen behind him at RL Day 2019.
    Philip Thomas stands in front of a projected screen at RL Day 2019 talking to about seated attendees.
    Poništi
  7. proslijedio/la je Tweet
    30. ruj 2019.

    Really clear and thoughtful review of an amazing decade of thinking by

    Poništi
  8. proslijedio/la je Tweet

    🎉We are pleased to announce the 3rd edition of the reproducibility challenge at ! 🎉 Researchers and students at all levels are encouraged to participate. Registration Deadline: Nov. 1 2019 Report Submission Deadline: Dec. 1 2019

    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    17. lip 2019.

    Is the Policy Gradient a Gradient?. Chris Nota and Philip S. Thomas

    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    16. lip 2019.
    Poništi
  11. proslijedio/la je Tweet
    5. lip 2019.

    Reinforcement Learning When All Actions are Not Always Available. Yash Chandak, Georgios Theocharous, Blossom Metevier, and Philip S. Thomas

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    5. lip 2019.

    Lifelong Learning with a Changing Action Set. Yash Chandak, Georgios Theocharous, Chris Nota, and Philip S. Thomas

    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet
    3. lip 2019.

    RL folks: are there any theoretical results for policy gradient methods that prove that lower variance via a baseline actually improves convergence rates (in certain settings)? All the arguments I've ever seen are super hand-wavy and basically just say "variance is bad".

    Poništi
  14. proslijedio/la je Tweet
    21. svi 2019.

    if the high-point of your @emnlp2019 submission is "we sota'ed" then consider arxiving instead, because your paper probably isn't very interesting for curious reviewers who want to learn (a.k.a. scientists who provide blind peer reviews as a service)

    Poništi
  15. proslijedio/la je Tweet
    10. svi 2019.
    Poništi
  16. proslijedio/la je Tweet
    14. tra 2019.

    Let's Play Again: Variability of Deep Reinforcement Learning Agents in Atari Environments. Kaleigh Clary, Emma Tosch, John Foley, and David Jensen

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet
    17. velj 2019.

    Reinforcement Learning Without Backpropagation or a Clock. James Kostas, Chris Nota, and Philip S. Thomas

    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    4. velj 2019.

    A Meta-MDP Approach to Exploration for Lifelong Reinforcement Learning. Francisco M. Garcia and Philip S. Thomas

    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    3. velj 2019.

    Learning Action Representations for Reinforcement Learning. Yash Chandak, Georgios Theocharous, James Kostas, Scott Jordan, and Philip S. Thomas

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    9. pro 2018.

    ToyBox: Better Atari Environments for Testing Reinforcement Learning Agents. John Foley, Emma Tosch, and Kaleigh Clary

    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:

    ·