Tushar Jain

@TusharJain_007

Research Engineer . Keep REINFORCE-ing concepts. MS Grad - Courant, NYU.

Vrijeme pridruživanja: lipanj 2012.

Tweetovi

Blokirali ste korisnika/cu @TusharJain_007

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @TusharJain_007

  1. Prikvačeni tweet
    8. pro 2018.

    Come and check out our ( ) paper on Multi-Agent Communication being presented during the Emmergent Communication Workshop 2018, NeurIPS. Contributed Talk: 9:45 am Poster: 12:40 pm Room: 524

    Poništi
  2. proslijedio/la je Tweet
    1. velj

    Curriculum for Reinforcement Learning "Learning is probably the best superpower we humans have." explores four types of curricula that have been used to help RL models learn to solve complicated tasks.

    Poništi
  3. proslijedio/la je Tweet
    26. sij

    10 hours fixing a bug = 5h looking for that bug + 1m fixing + 4h59m complaining to everyone about it

    Poništi
  4. proslijedio/la je Tweet
    29. sij

    Our paper 'A distributional code for value in dopamine-based reinforcement learning' on the cover of ! Read it here: Shout out to the amazing artists/designers at who make this possible, while we get to focus on the research.

    Poništi
  5. proslijedio/la je Tweet
    30. sij

    Excited to give a talk today on ANML: Learning to Continually Learn. Recording available soon. If you are curious about ANML now, here is a NeurIPS talk I gave that covers it (amongst other things). Talk today is at the SF summit

    Poništi
  6. proslijedio/la je Tweet
    20. sij

    I analyzed compensation & level details of 19k tech workers to find answers to: 1. How long does it take for SWEs to reach a certain level? 2. Compensations across jobs/levels? 3. Do women get paid less than men in tech? 4. Is there a deadline for SWEs?

    Poništi
  7. proslijedio/la je Tweet

    This is extraordinarily important. In order to evaluate a team and its management, see who is getting the stretch opportunities. I personally learned more from stretch opportunities than anything else.

    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    5. stu 2019.

    Thought-provoking abstract by - pioneer of dependency parsing - on whether the end of parsing is in sight. Admirably honest, but also a little sad... It refers back to a polemic paper of my Amsterdam colleague, and former boss, 1/n

    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    16. sij

    If you run a team of researchers you are responsible for the culture. You can set the trend, encourage good practice, address bullying, intimidation, change the reward system. Passive ignorance is not acceptable.

    Poništi
  10. proslijedio/la je Tweet
    17. sij

    Reformer: The Efficient Transformer Blog Post “The Reformer is a Transformer model designed to handle context windows of up to 1 million words, all on a single accelerator and using only 16GB of memory.”

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

    Amazing! This real working deep learning image classifier is running, for free, on , is written entirely in notebooks with ipywidgets, & deployed with Voila. 13 lines of code :) Try it here:

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    7. sij

    You've read the paper (all three and a half of you, thanks!) Now see it live this Friday at BayesComp 2020! Room C8 at 11:20am.

    Poništi
  13. proslijedio/la je Tweet
    4. sij

    This is an excellent exploration of neural style transfer with Weights & Biases. We show you how a neural network learns to separate an image's style & content; & how to iterate your way to the best model.

    Poništi
  14. proslijedio/la je Tweet
    4. sij

    We shouldn't be afraid to develop benchmarks based on goals. It still allows us to explore, like in disentanglement research, the full spectrum from fully-supervised to unsupervised discovery of goals. Particularly inspiring is NLP models conditioned to perform very diff. tasks.

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

    Today is my last day at Uber. It's been a dream to have our startup acquired by Uber &, with wonderful colleagues, create Uber AI Labs. I'm proud of our work, eg. Go-Explore, POET, AI-GAs, GTNs, ANML, Differentiable Plasticity, Backpropamine, Deep Neuroevolution, PPGNs, etc. 1/3

    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    4. sij

    I want to deeply thank the University of Wyoming and all the wonderful people there, especially my fantastic students. I will not be returning, but it was an excellent environment for research and teaching. I am proud of the lab we created and its work 1/3

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

    TIL never to merge to master late on Friday afternoon ... even when the unit tests pass ...

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

    My last 3650 days: 👨‍👧‍👧 Kids: 1 → 2 👩‍❤️‍👨 Marriage: 3 → 13yrs 👨‍👧‍👦 Nieces/nephews: 3 → 10 🏢 Webflow: 1 → 150ppl 💰 Customers: ~25 → 100K+ 🐦 Followers: 0 → 25K 💵 Loans: ~$80K → 0 👵 Grandparents: 4 → 1 :( Reminds me... “The days are long, but the years are short.” —G. Rubin

    Poništi
  19. proslijedio/la je Tweet
    30. pro 2019.

    If you already went through the tough work of publishing twice at peer-reviewed venues, and really just after a PhD rather than the process, have you thought of going for a thesis-only PhD, rather than spend years working for a PI? Some places offer these

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    25. pro 2019.
    Prikaži ovu nit
    Poništi
  21. proslijedio/la je Tweet
    27. pro 2019.

    Reasons I love writing and think everyone should write more: 1. To organize your thoughts 2. To learn: the best way to learn is to write/teach 3. To keep you accountable 4. To document your progress 5. To write better 6. To put your name out there 7. To enjoy the beauty of words

    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:

    ·