OpenAIVerified account

@OpenAI

OpenAI’s mission is to ensure that artificial general intelligence benefits all of humanity. We're hiring:

Joined December 2015

Tweets

You blocked @OpenAI

Are you sure you want to view these Tweets? Viewing Tweets won't unblock @OpenAI

  1. Jan 30

    We're standardizing OpenAI's deep learning framework on PyTorch to increase our research productivity at scale on GPUs (and have just released a PyTorch version of Spinning Up in Deep RL):

    Undo
  2. 13 Dec 2019

    We're releasing "Dota 2 with Large Scale Deep Reinforcement Learning", a scientific paper analyzing our findings from our 3-year Dota project: One highlight — we trained a new agent, Rerun, which has a 98% win rate vs the version that beat .

    Undo
  3. 5 Dec 2019

    A surprising deep learning mystery: Contrary to conventional wisdom, performance of unregularized CNNs, ResNets, and transformers is non-monotonic: improves, then gets worse, then improves again with increasing model size, data size, or training time.

    Undo
  4. 3 Dec 2019

    We're releasing Procgen Benchmark, 16 procedurally-generated environments for measuring how quickly a reinforcement learning agent learns generalizable skills. This has become the standard research platform used by the OpenAI RL team:

    Undo
  5. 21 Nov 2019

    We're releasing Safety Gym, environments and tools to evaluate reinforcement learning with safety constraints: Aims to ultimately help agents satisfy real-world safety requirements while training (eg not driving off a cliff, not writing abusive content).

    Undo
  6. 7 Nov 2019

    We've analyzed compute used in major AI results for the past decades and identified two eras in AI: 1) Prior to 2012 - AI results closely tracked Moore's Law, w/ compute doubling every two years. 2) Post-2012 - compute has been doubling every 3.4 months

    Show this thread
    Undo
  7. 5 Nov 2019

    We're releasing the 1.5billion parameter GPT-2 model as part of our staged release publication strategy. - GPT-2 output detection model: - Research from partners on potential malicious uses: - More details:

    Undo
  8. 16 Oct 2019

    In case you missed it, here’s the unedited solve of the Rubik’s cube:

    Show this thread
    Undo
  9. 16 Oct 2019

    Human hands let us solve a wide variety of tasks. Even so, solving a Rubik's Cube one-handed isn't easy for humans. We're excited to continue to develop new AI technology and ultimately ensure that these systems benefit all of humanity.

    Show this thread
    Undo
  10. 16 Oct 2019

    "Solving the Rubik's Cube with a Robot Hand" took many human hands over the past 2.5 years — meet our Robotics team! (PS they're hiring: !)

    Undo
  11. 15 Oct 2019

    We’re all used to robots that fail when their environment changes unpredictably. Our robotic system is adaptable enough to handle unexpected situations not seen during training, such as being prodded by a stuffed giraffe:

    Show this thread
    Undo
  12. 15 Oct 2019

    We've trained an AI system to solve the Rubik's Cube with a human-like robot hand. This is an unprecedented level of dexterity for a robot, and is hard even for humans to do. The system trains in an imperfect simulation and quickly adapts to reality:

    Show this thread
    Undo
  13. 11 Oct 2019

    Now accepting applications for our 3rd class of OpenAI Scholars: a 4 month full-time program for individuals from underrepresented groups to study deep learning and produce an open-source project. Mentors include , , :

    Undo
  14. 19 Sep 2019

    Wondering why the hiders did not cage in the seekers instead of building their own fort? In one environment variant where hiders have to protect glowing orbs, that's exactly what they learned to do!

    Undo
  15. 19 Sep 2019

    We've fine-tuned GPT-2 using human feedback for tasks such as summarizing articles, matching the preferences of human labelers (if not always our own). We're hoping this brings safety methods closer to machines learning values by talking with humans.

    Undo
  16. 17 Sep 2019

    And seekers learn that if they run at a wall with a ramp at the right angle, they can launch themselves upward.

    Show this thread
    Undo
  17. 17 Sep 2019

    Unexpected and surprising behaviors included box surfing, where seekers learn to bring a box to a locked ramp in order to jump on top of the box and then “surf” it to the hider’s shelter.

    Show this thread
    Undo
  18. 17 Sep 2019

    We've observed AIs discovering complex tool use while competing in a simple game of hide-and-seek. They develop a series of six distinct strategies and counterstrategies, ultimately using tools in the environment to break our simulated physics:

    Show this thread
    Undo
  19. 27 Aug 2019

    . recently chatted with in his latest podcast, "Behind the Tech," and shared his thoughts on AI progress, building a mission-driven company, and the future of transformative technologies:

    Undo
  20. 22 Aug 2019

    We're releasing a new method to test for model robustness against adversaries not seen during training, and open-sourcing a new metric, UAR (Unforeseen Attack Robustness), which measures how robust a model is to an unanticipated attack:

    Undo

Loading seems to be taking a while.

Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

    You may also like

    ·