Tweetovi

Blokirali ste korisnika/cu @sidgreddy

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

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

    How do you train an RL agent in the presence of unknown, unsafe states without visiting them even once? New algorithm by our intern synthesizes trajectories with a generative model and ask a human to label them for safety.

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    13. pro 2019.

    We present ReQueST: a method for training RL agents from human feedback in the presence of unknown unsafe states. By , , , , Paper: Code:

    Prikaži ovu nit
    Poništi
  3. 7. lis 2019.

    Scaled Autonomy: Enabling Human Operators to Control Robot Fleets Imitation learning for mimicking expert actions, and imitation learning for deciding which robot the expert should teleop right now. w/, ,

    Poništi
  4. proslijedio/la je Tweet
    28. svi 2019.

    SQIL: an imitation learning method so simple I can summarize in a tweet: drop demonstrations into buffer, set their reward to +1, set reward for all other data to 0, run Q-learning or SAC to train. Why does it work? Find out: w/

    Poništi
  5. proslijedio/la je Tweet
    23. svi 2018.

    Where do You Think You’re Going? New paper on inferring how humans believe an MDP (e.g., game) works With @elbowproxy and A. Dragan

    Poništi
  6. proslijedio/la je Tweet
    21. svi 2018.

    "Where Do You Think You're Going?: Inferring Beliefs about Dynamics from Behavior," Reddy et al.:

    Poništi
  7. proslijedio/la je Tweet
    18. tra 2018.

    Interested in semi-autonomous robots and human-machine centaur teams? Read @elbowproxy’s new BAIR blog post on augmenting human control with deep reinforcement learning for flexible human-robot shared autonomy!

    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:

    ·