Tweetovi

Blokirali ste korisnika/cu @AngeloDalli

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

  1. 2. velj

    Today is a rare : 02 02 2020 It’s also and also the 33rd day of the year with 333 days left - maybe we’ll soon go into a time loop like Bill Murray in the 90s :) The last palindrome date was 909 years ago on 11/11/1111, the next on 12/12/2121!

    Poništi
  2. 30. sij

    For those interested in understanding cause and effect have a look at new article summary page!

    Poništi
  3. 29. sij

    An international panel discussion about ‘How will A.I shape an evolving job market?’ & Understanding the paradigm shift in the job market. The panel included Jorge Sebastio, & myself.

    Poništi
  4. 28. sij

    Coronavirus patient being treated by a robot - truly useful application of and points to the future of medical care esp in critical cases

    Poništi
  5. 21. sij

    My interview with Coin Geek at the 2019! “We are seeing more and more practical technologies affecting all industries such as health care and payments and making it easier for everyone to use it....” !

    Poništi
  6. 16. sij

    After around thirty years learning from and knowing Ġorġ Mallia I finally have my own cartoon sketch! Am happy to announce that 111 Art will be hosting an exhibition of his works on 6th March - more details to follow soon!

    Poništi
  7. 13. sij

    Check out my latest interview where I am discussing - what the real uses of AI are, my thoughts on merging regulation & AI and whether AI can help humanity make tough decisions. Let me know your thoughts on these advancing factors.

    Poništi
  8. 9. sij

    Happy to have been in the ‘Top 10 most viewed interviews of the decade’ for technology, artificial intelligence, igaming and more on Calvin Ayre with Becky Fontana. This took place in 2014, check it out here!

    Poništi
  9. 8. sij

    What is Umnai? Check out this short interview about Umnai and its impact on the Artificial Intelligence field!

    Poništi
  10. 7. sij

    Business Angels Malta Launch featured on ’s website! Check it out here

    Poništi
  11. 6. sij

    that can analyse satellite imagery is being restricted for export by the US. Do you think this overly restrictive or completely justified?

    Poništi
  12. 3. sij

    Awesome result from showing that human neurons can compute XOR - in plain English: proof that our brains are effective computers :) Not seen in other species! Huge boost for AI neural models!  

    Poništi
  13. 31. pro 2019.

    Balanced discussion between & that makes sense. Hope that in 2020 ppl will help find the strengths and limits of different tools available and combine them in a superior way for future AI.

    Poništi
  14. 29. pro 2019.

    Missing in is EXPLAINABILITY - DL or some hybrid future needs to explain itself. How else would an AI system prove that it actually understands whatever it’s being trained on? objectives could be priors?

    Poništi
  15. 28. pro 2019.

    2/2 some things like Causal Logic, symbolic logic, rule systems are not part of DL. I strongly believe joining these up will give the best results. Swallow pride on all sides and call it something else :) Recommend for independent opinion

    Prikaži ovu nit
    Poništi
  16. 28. pro 2019.

    Interesting redefinition of Deep Learning by Yoshua Bengio - emphasis on gradient descent methods is out and now DL encompasses all neural networks... there’s a need to classify types of systems to know their limitations! continues!

    Poništi
  17. 26. pro 2019.

    Key takeaway from and Bengio : systems need to have inbuilt prior knowledge like basic physics, inferencing, cause & effect, explanation capabilities, etc to go beyond current DL limits. Then using some future algorithm to generalize, learn & understand

    Poništi
  18. 23. pro 2019.

    Bengio’s version of Occam’s razor: Consistent generative future AI systems should *start off with the least number of priors* — would these minimal priors be the start of a causal graph that can handle most common sense situations? Do findings hold?

    Poništi
  19. 23. pro 2019.

    “We need to avoid pitfalls of classical AI rule based symbol manipulation, with changes handled using causality to allow deep learning to go beyond training data and generalize. Efficient search & neural net symbol reps allow for richer processing” - Yoshua Bengio

    Poništi
  20. 23. pro 2019.

    “Going forward researchers should be working on compositionality, reasoning, causality, hybrid models, and extrapolation beyond the training space”

    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:

    ·