John O'Malia

@john_omalia

Transformative systems merging human insight and context with AI computation:

Vrijeme pridruživanja: listopad 2016.

Tweetovi

Blokirali ste korisnika/cu @john_omalia

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

  1. 18. sij

    For anyone considering how to make deep learning more versatile, read this and consider using more semantic / conceptual mappings in your learning systems! My spin - break with tradition and explicitly embrace human biases for improved disentanglement in DRL!!

    Poništi
  2. proslijedio/la je Tweet
    16. sij

    Looking forward to sharing our lessons learned building voice-controlled characters in games London! We'll share some amazing human-AI teaming insights from Monday at 3PM in the Big Screen Gaming track!

    Poništi
  3. 20. pro 2019.

    Incredibly proud to apply human-directed AI technology to EasyAerial to bring voice controlled drones to first responders and perimeter security on this USAF SBIR!

    Poništi
  4. 15. stu 2019.

    So looking forward to talking about voice-controlled characters, player/AI teaming in games, and our upcoming product launch at !! I hope to see you there in January!

    Poništi
  5. 18. lis 2019.

    Hugely proud of what the team has done to drop human-directed characters into games! The full stack works within the Unity game engine, so any game can turn on voice-controlled characters with an easy integration. The future is human/AI teaming 🚀 !!

    Poništi
  6. 11. lis 2019.

    Great to see more folks working on combining human instructions using natural language with AI agent action selection! Human/AI teaming is the future...Watch this space 😀

    Poništi
  7. proslijedio/la je Tweet
    10. lis 2019.

    DM set out their stall on Graph Networks with their original Relational Inductive Biases paper. This work makes clear how powerful graph networks can be when we integrate inductive biases into GN's. Much more to come in this direction from them I'm sure (and...from us).

    Poništi
  8. 8. lis 2019.

    Yes - we're with you on this approach. Improving neural network performance by giving them access to causal knowledge, whether they derive it or learn it adds a ton of value and we're leveraging this in other ways.

    Poništi
  9. 29. ruj 2019.

    The student/teacher model applied here for language model compression with shared projection matrices reminds me so much of the interplay between hippocampus and cortex. Shout out to and the BrainInspired podcast for prompting thoughts of crossover

    Poništi
  10. 4. ruj 2019.

    This is a game changing, carbon-capturing solution for the US insulation and construction markets. And it's a big boost for Maine. So proud of my brother!

    Poništi
  11. 20. kol 2019.

    The thesis: Combining human intuition with AI computation will drive the next generation of breakthroughs in art, science, drug discovery, etc. As so often happens, artists are showing the way - here exploring latent space embeddings and GANs.

    Poništi
  12. 14. kol 2019.

    Brilliant post... Anyone interested in how people actually interact in chat, and what engages them in terms of bot structure, is going to learn something here. I love the insight that bots can be clearly non-human yet highly engaging...

    Poništi
  13. 4. kol 2019.

    Dileep is correct. Our view is that only when you can explicitly reuse and extend information/knowledge in new contexts can you (eventfully!) find a path to AGI.

    Poništi
  14. 19. srp 2019.

    We love having more people join us in promoting Human AI Teaming! 🔥 initiative, and although we are much more learning focused than this approach, notes correctly that it's a great baseline. will soon release some of our H/AI work too 👍😀

    Poništi
  15. 16. srp 2019.

    Random homeschooling data point regarding instilling a love of reading (he's 7): Yesterday, visit to the library: got 6 books. He finished 2 (light, ~100p) that day,, started 3rd. Today: finished the 3rd (Trials of Apollo, Riordan) 374p Read (tons) with your kid! It works!

    Poništi
  16. 11. srp 2019.

    Good call from Human/AI teaming is the way forward... Only when we combine human insight with AI computation will we really see what's possible. Some exciting stuff coming from in this space!

    Poništi
  17. 4. srp 2019.

    What terrific work! We live in remarkable times, to have things like this popping up. This presents the AI field with some fascinating food for thought regarding symbols and abstraction, not to mention reward modeling. The mapping of experience to other vectors is fascinating

    Poništi
  18. 27. lip 2019.

    Huge congratulations to our amazing team at , who have just won EGR's B2B Software Rising Star Award! Turning AI theory into robust, high-value-add SAAS products is equally challenging and rewarding., but when it works, the effects are near-magical.

    Poništi
  19. 6. lip 2019.

    Honored to participate with in the yesterday. Deeply impressed by the and teams' commitment to ethically sound approaches to AI which advance the safety and security of the UK.

    Poništi
  20. 15. svi 2019.
    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:

    ·