Sanuj

@sanuj_sharma

ML/NLP grad student. Ex-full stack engineer 📜 👨‍💻 🎨 Prev: [YC W17], ,

Vrijeme pridruživanja: svibanj 2015.

Tweetovi

Blokirali ste korisnika/cu @sanuj_sharma

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

  1. proslijedio/la je Tweet
    28. ruj 2019.

    11 years ago today, we launched our first successful mission. To date, we’ve completed 78 launches and have developed the world’s only operational reusable orbital class rockets and spacecraft—capable of launching to space, returning to Earth, and flying again

    Poništi
  2. proslijedio/la je Tweet

    We're having an event next Tuesday in San Francisco to share a bit about what we've been working on the last two years, and we’ve reserved a few seats for the internet. Apply here:

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

    🚀 Excited to share with you all: ! It’s a website which helps you select books to read. 📚

    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet

    I used autopilot 99% of the drive. I would go over 100 miles and over an hour of driving without providing any input to the car. no steering, no breaking no accelerating - just autopilot. it's way better than it gets credit for.

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet

    Next up in 2A Dialog and Discourse: Sanuj Sharma from talking about dialog state tracking from context.

    Prikaži ovu nit
    Poništi
  6. proslijedio/la je Tweet
    23. svi 2019.

    What unites people? Armies? Gold? Flags? Stories. There’s nothing in the world more powerful than a good story. Nothing can stop it. No enemy can defeat it. – Tyrion Lannister, on how to write a good paper

    Prikaži ovu nit
    Poništi
  7. 3. svi 2019.

    Good survey paper on Episodic Meta RL by Matthew Botvinick et al. (Deepmind) Highlights possible solutions to current limitations in deep RL and some parallels with neuroscience and psychology. Talk: Reddit:

    Poništi
  8. 26. tra 2019.
    Poništi
  9. proslijedio/la je Tweet
    26. tra 2019.

    A neat work that improves upon our GLAD model by incorporating relevant dialogue context to obtain a new SOTA in dialogue state tracking! By out of Ruihong Huang's group at Texas A&M

    Poništi
  10. proslijedio/la je Tweet
    25. tra 2019.

    I saw this video of Boston Dynamics' testing facility 5 days ago and I'm still thinking about it.

    Prikaži ovu nit
    Poništi
  11. 22. tra 2019.

    Key ideas behind auto-ML / auto-keras: 1. Bayesian Optimization. () 2. Network Morphism. ()

    Poništi
  12. proslijedio/la je Tweet

    In the very long run, it's going to be 0K.

    Poništi
  13. proslijedio/la je Tweet
    11. tra 2019.

    It is hard to express how utterly remarkable and astoundingly delightful it is that mathematical scribblings can predict something as bizarre as black holes and then, looking to the skies with sufficient diligence, we find them.

    Poništi
  14. proslijedio/la je Tweet

    We asked and checked. The space telescope concept can indeed fly on Starship! (graphic used by permission)

    Graphic illustrating LUVOIR space telescope folded into body of SpaceX's Starship
    Poništi
  15. proslijedio/la je Tweet
    8. tra 2019.

    I asked how to set neural network init. He accidentally replied with a poem: You want to be on the edge of chaos Too small, and the init will be too stable, with vanishing gradients Too large, and you'll be unstable, due to exploding gradients You want to be on the edge

    Poništi
  16. 7. tra 2019.

    My first paper in NAACL: Improving Dialogue State Tracking by Discerning the Relevant Context LSTM based architecture to model dialog state with simple dialog history lookup and gated user/system utterance fusion.

    Poništi
  17. proslijedio/la je Tweet
    5. tra 2019.
    Poništi
  18. 6. tra 2019.
    Poništi
  19. proslijedio/la je Tweet
    27. ožu 2019.

    The X factor: When I was an undergrad at Kings College Cambridge, Les Valiant who won the Turing award in 2010 lived in the adjacent room on X staircase. He just told me that Turing lived on X staircase when he was a fellow at Kings and probably wrote his 1936 paper there!

    Poništi
  20. proslijedio/la je Tweet
    1. ožu 2019.

    Different parts of the brain learns in different ways. Model-free, model-based, and memory-based are all just different sides of the same coin, according to Kenji Doya

    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:

    ·