Miles Brundage

@Miles_Brundage

Research Scientist (Policy) at OpenAI. Research Associate, Future of Humanity Institute, University of Oxford. Views my own.

San Francisco, CA
milesbrundage.com
Vrijeme pridruživanja: veljača 2013.
1.419 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @Miles_Brundage

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

  1. Prikvačeni tweet
    20. kol 2019.

    Really excited to be sharing the results of our + others' research into language models and their social impacts, as well as what we've learned on publication norms in the past 6 months. Direct link to the report:

    GPT-2 6-month follow-up: we're releasing the 774M parameter model, an open-source legal doc organizations can use to form model-sharing partnerships, and a technical report about our experience coordinating to form new publication norms:
    34 proslijeđena tweeta 134 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    prije 9 sati

    Are AI DJs as good as human ones? Not remotely, as 's article notes. They're just cheaper. A bunch of radio stations just got shoddier, and iheartradio's executives just got $33 million in bonuses.

    15 proslijeđenih tweetova 30 korisnika označava da im se sviđa
    Poništi
  3. proslijedio/la je Tweet
    1. velj

    On challenges of learning human preferences...

    How do you teach an system to understand humans' preferences if humans aren't always sure what they want? HAI affiliates , Stuart Russell and Yoshua Bengio are exploring the answers.
    8 proslijeđenih tweetova 46 korisnika označava da im se sviđa
    Poništi
  4. proslijedio/la je Tweet
    31. sij

    Again, the 'slaughterbots' in this video are semi-AWS - their targets were selected by human beings. THEY WOULDN'T BE PROHIBITED BY A BAN ON FULLY AUTONOMOUS WEAPON SYSTEMS. Maybe we should focus on regulating the use of semi-AWS instead of misleading scaremongering.

    We should lead the world in a global ban on lethal autonomous weapons systems. Only humans should be making life or death decisions.
    4 proslijeđena tweeta 23 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    30. sij

    Check it out: Spinning Up is now a sort of "Rosetta Stone" for deep RL, with simple implementations of VPG, PPO, DDPG, TD3, and SAC in both PyTorch and TFv1.

    16 proslijeđenih tweetova 87 korisnika označava da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    30. sij

    The new acronym for FAT* is here! Starting with the 2021 conference, we’ll be ACM FAccT for ‘Fairness, ACCountability, and Transparency’ 1/

    7 replies 67 proslijeđenih tweetova 137 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    30. sij

    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):

    25 replies 545 proslijeđenih tweetova 1.916 korisnika označava da im se sviđa
    Poništi
  8. proslijedio/la je Tweet
    29. sij

    Facebook is paying $550 million to settle a class action lawsuit brought in Illinois, where residents are protected by the nation's strongest biometric privacy law.

    6 replies 123 proslijeđena tweeta 288 korisnika označava da im se sviđa
    Poništi
  9. proslijedio/la je Tweet
    28. sij

    Introducing , a 2.6B-param open-domain chatbot with near-human quality. Remarkably, we show strong correlation between perplexity & humanlikeness! Paper: Sample conversations:

    i
    14 replies 92 proslijeđena tweeta 339 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    28. sij

    Even against the backdrop of strategic distrust, major militaries should cooperate on measures to improve the safety, surety, and security of AI systems in military affairs, and I argue in this commentary for via .

    1 reply 16 proslijeđenih tweetova 28 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    28. sij

    In “Artificial Intelligence, Values and Alignment” DeepMind’s explores approaches to aligning AI with a wide range of human values:

    6 replies 147 proslijeđenih tweetova 374 korisnika označavaju da im se sviđa
    Poništi
  12. proslijedio/la je Tweet
    27. sij

    Launching – Our journey bringing together leading organisations to improve understanding of Artificial Intelligence -

    1 reply 12 proslijeđenih tweetova 12 korisnika označava da im se sviđa
    Poništi
  13. proslijedio/la je Tweet
    27. sij

    A lot has been said about the dangers of large-scale language model trained on the internet, like GPT-2. We fine-tuned a version of GPT-2 to avoid generating descriptions of non-normative behavior (killing, suicide, things inconsistent with social norms)

    14 replies 68 proslijeđenih tweetova 266 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    24. sij

    Class-action lawsuit filed against controversial Clearview AI startup

    101 proslijeđeni tweet 174 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    23. sij

    Nice summary of and my paper on AI Ethics and Policy research priorities. Looking forward to presenting it at in February.

    Cool short paper on the false dichotomy between 'near-term' & 'long-term' AI safety considerations. Authors suggest a more helpful model with four different dimensions: capabilities, impacts, certainty and extremity, which all sit on a spectrum.
    4 proslijeđena tweeta 18 korisnika označava da im se sviđa
    Poništi
  16. proslijedio/la je Tweet
    20. sij

    I started a weekly email newsletter for people interested in seeing the most interesting machine learning conversations on Twitter. Each issue highlights 5-10 arXiv papers + related tweets. Check out the latest issue here and subscribe to get it next week!

    1 reply 22 proslijeđena tweeta 107 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet
    20. sij

    Submit your datasets, papers, + abstracts to the 1st-ever AI for Earth Sciences workshop ! 🌍 We're looking for AI/ML applied to challenges in atmospheric physics, oceanography, geology, planetary sciences, space weather, + geo-health:

    31 proslijeđeni tweet 70 korisnika označava da im se sviđa
    Poništi
  18. proslijedio/la je Tweet
    20. sij

    Given the smoothness of videos, can we learn models more efficiently than with ? We present Sideways - a step towards a high-throughput, approximate backprop that considers the one-way direction of time and pipelines forward and backward passes.

    7 replies 268 proslijeđenih tweetova 1.012 korisnika označava da im se sviđa
    Poništi
  19. 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?

    285 proslijeđenih tweetova 1.129 korisnika označava da im se sviđa
    Poništi
  20. proslijedio/la je Tweet
    18. sij

    The privacy paranoid among us have long worried that all of our online photos would be scraped to create a universal face recognition app. My friends, it happened and it’s here:

    389 replies 10.501 proslijeđeni tweet 13.193 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  21. proslijedio/la je Tweet
    18. sij

    Great to see my summary of the discussions at the third Economics of AI conference in . To sum up, it is hard to sum up. The impact of AI on the economy will be complicated.

    From to , many worry about the significant economic future impacts of AI. Economist predicts that instead, “the future of AI in the economy will resemble the Internet more than Skynet.”
    8 proslijeđenih tweetova 19 korisnika označava da im se sviđa
    Poništi

@Miles_Brundage još nije objavio nijedan Tweet.

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:

    ·