Joel Lehman

@joelbot3000

Senior Research Scientist @ Uber AI Labs, Lead on AI Safety & Ethics

Vrijeme pridruživanja: lipanj 2009.

Tweetovi

Blokirali ste korisnika/cu @joelbot3000

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @joelbot3000

  1. proslijedio/la je Tweet
    13. pro 2019.

    There's definitely a common philosophical thread running through 's latest essay and and Joel Lehman's "Why Greatness Cannot Be Planned"

    Poništi
  2. proslijedio/la je Tweet
    18. pro 2019.

    Introducing Generative Teaching Networks - they generate synthetic data optimized to train learners as fast as possible - a new way to generate data for architecture search & more! Led by w/ , , &

    Poništi
  3. 12. pro 2019.

    Have you ever seen an inchworm crawl up a leaf or twig and there, clinging to the very end, revolve in the air feeling for something to reach? That's like me. I'm trying to find something out there beyond the place I have a footing. -Albert Pinkham Ryder

    Poništi
  4. proslijedio/la je Tweet
    9. pro 2019.

    1/13 How to revolutionize AI and become famous. First consider a) through e), then follow the instructions.

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet

    Curious what thinks about the Neural ODE paper () in retrospect? Come see his talk at the Retrospectives workshop next Friday!! Full schedule here:

    Poništi
  6. proslijedio/la je Tweet
    12. pro 2019.

    If you are willing to learn more about Quality-Diversity algorithms, we have created a website that gathers papers, tutorials, and implementations of QD algorithms. The progressive increase of papers published on this topic every year is impressive!

    Prikaži ovu nit
    Poništi
  7. 10. pro 2019.

    Melanie's seminal copycat work is super interesting, curious to hear how she views it now @ Friday's retrospectives workshop

    Poništi
  8. proslijedio/la je Tweet

    We've decided to extend the workshop deadline one last time until after ICLR: Sunday, Sep 29!! This will be the last extension, since author notifications are required by Oct 1. Note that accepted papers will get guaranteed tickets! 👌

    Poništi
  9. proslijedio/la je Tweet

    Announcement: we're extending the deadline for the Retrospectives workshop by 1 week, until Sunday Sep 22!! Take an hour to write up some thoughts about a past paper of yours (they don't need to be critical!) and submit by creating an issue on 🧐

    Poništi
  10. proslijedio/la je Tweet
    12. ruj 2019.

    Don't aim for success if you want it; just do what you love and believe in, and it will come naturally. - David Frost (via the excellent book Why Greatness Cannot be Planned by and ).

    Poništi
  11. proslijedio/la je Tweet
    10. ruj 2019.

    Reminder: the deadline for the NeurIPS Retrospectives workshop is in less than a week (this Sunday)!! Also, it's not just for retrospectives: you can submit meta-analysis papers too! (analyze a set of papers that aren't your own) 🙂

    Poništi
  12. proslijedio/la je Tweet
    10. ruj 2019.

    Surprising neural net properties via simple "LCA" measure by Uber AI Resident with + ! 👉 barely over half of params help vs. hurt each iter 👉 some layers move against grad overall(!) 👉 micro-learning synced across layers 👇read more on blog👇

    Poništi
  13. proslijedio/la je Tweet

    Does science self correct such that non-replicating papers get cited less? Apparently not. The prediction market can predict what won't replicate, but apparently the field doesn't.

    Poništi
  14. 8. ruj 2019.

    'I hate crowds in science ... I tell my students, when you go to these meetings, see what direction everyone is headed, so you can go in opposite direction. Don't polish the brass on the bandwagon' -VS Ramachandran on big conferences c/o The Brain that Changes Itself

    Poništi
  15. proslijedio/la je Tweet

    Hello, world! ML Retrospectives is officially live!! We're a website that hosts retrospectives: blog posts where researchers talk openly about their past papers. Check it out at . It comes with a workshop at 2019!

    Poništi
  16. 21. kol 2019.

    "It’s not uncommon to read a paper in ML and think: 'What secret trick do you need to actually get this to work?' ... We’ve grown accustomed ... to such a large amount of paper debt that this second guessing seems normal." Great piece by re: upcoming NeurIPS workshop

    Poništi
  17. 13. kol 2019.
    Poništi
  18. 1. kol 2019.

    Making art out of political dysfunction

    Poništi
  19. proslijedio/la je Tweet

    Uber's EvoGrad is a dev library for evolutionary algorithms by

    Poništi
  20. proslijedio/la je Tweet
    22. srp 2019.

    Announcing our EvoGrad library for grad-based evolution + our Evolvability ES meta-learning algorithm: can scale to deep nets, compete w/ MAML in RL. With , , and . Blog: , paper:

    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:

    ·