Bradley Love

@ProfData

Fellow at Alan Turing Institute for data science and ; Professor of Cognitive and Decision Sciences, University College London

London, UK
Vrijeme pridruživanja: rujan 2014.

Tweetovi

Blokirali ste korisnika/cu @ProfData

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

  1. Prikvačeni tweet
    2. pro 2019.

    New preprint, "Levels of Biological Plausibility", covering how understanding of mechanism, reduction, and emergence are tied to levels, as well as how biological plausibility is an incoherent concept under a levels of mechanism view.

    Poništi
  2. proslijedio/la je Tweet
    27. sij

    Our new preprint just out from looking at the 3.9 million participants in "Cities have a negative impact on navigation ability: evidence from 38 countries" Work led by the fantastic My 1st thread: 1/n

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    28. sij

    Our take on the new paper by & - covers why we think alignment of different conceptual systems is important. Written with dev psych . Journal link here and free text link in next tweet...

    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    27. sij
    Poništi
  5. proslijedio/la je Tweet
    23. sij

    An Existential Crisis in Neuroscience The hopes that ever more detailed brain mapping would make function easier to understand are fading

    Poništi
  6. 20. sij

    "Learning as the Unsupervised Alignment of Conceptual Systems" now has a view-only version from the publisher (see below) in addition to It's also made it to sci-hub for those so inclined.

    Poništi
  7. proslijedio/la je Tweet
    19. sij

    Working with large data sets is a kind of labor that is too often left out of scientific training. Here are 11 tips for making the most of your large data sets.

    Poništi
  8. proslijedio/la je Tweet
    19. sij

    tl;dr communicative codes don't work like you think they do.

    Poništi
  9. proslijedio/la je Tweet
    17. sij

    Very nice paper! Multi-modal alignment allows for unsupervised discovery of concepts/labels/meaning. IMO these ideas will eventually allow language models to exhibit deep "understanding" of text, and will solve the "lack of common sense" problem

    Poništi
  10. 17. sij

    Oh, if your institution doesn't have access, it's all here,

    Prikaži ovu nit
    Poništi
  11. 17. sij

    We've shown that supervised learning can be solved by purely unsupervised means. The big challenge now is to devise algorithms to efficiently exploit this information across conceptual systems. We (and I hope others) are working to meet this challenge. 5/5

    Prikaži ovu nit
    Poništi
  12. 17. sij

    From a developmental perspective, children's early concepts form readily aligned systems, which in principle allow for purely unsupervised learning without image-label co-occurrence, child directed speech, etc. 4/5

    Prikaži ovu nit
    Poništi
  13. 17. sij

    Unsupervised alignment only gets better with more concepts and systems. Each concept's signature becomes more unique the more concepts there are in each system (see below). Each system provides a slightly different viewpoint on reality such that adding more systems helps too. 3/5

    Prikaži ovu nit
    Poništi
  14. 17. sij

    How can supervised tasks be solved without supervision? Concepts have unique signatures that hold across unsupervised systems. E.g., The words pen and pencil pattern similarly in text corpora and are also visually similar. Likewise, both are different to lions, cars, etc. 2/5

    Prikaži ovu nit
    Poništi
  15. 17. sij

    New paper w , "Learning as the Unsupervised Alignment of Conceptual Systems". Supervised learning tasks can be solved by purely unsupervised means by exploiting correspondences across systems (e.g., text, images, etc.). 1/5

    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    14. sij

    The Turing is now welcoming applications for our Data Science for Social Good Summer Fellowship. This is a 12-week programme run in partnership with and . Read more and apply➡️

    Poništi
  17. proslijedio/la je Tweet
    7. sij

    Very excited about our new paper: "Ventromedial prefrontal cortex compression during learning". vmPFC coding akin to goal-directed dimensionality reduction and tracks attention weights from learning models.

    Poništi
  18. 7. sij

    Out today, "Ventromedial prefrontal cortex compression during concept learning" w . vmPFC compresses information during learning in a goal-directed manner. Its timecourse is unique and parallels attention weights from learning models.

    Poništi
  19. proslijedio/la je Tweet
    7. sij

    Interested in doing a PhD on the computational neuroscience of mental health? Only few more days to apply for this fantastic PhD programme: If you fancy me as your supervisor - please get in touch.

    Poništi
  20. proslijedio/la je Tweet

    Harsh but fair: “Unfortunately, what these discussions demonstrate is that many researchers do not, as a rule, actually understand the formal definitions of ‘computer’ or ‘algorithm’ as provided by computer science.” — 

    Poništi
  21. proslijedio/la je Tweet
    16. pro 2019.

    Our theme issue is out: If you are interested in compositionality from formal, computational, or neurophysiological/neurobiological perspectives with a mechanistic bent, eat your heart out! It was an honour to edit this together with the inimitable Giosuè Baggio .

    Prikaži ovu nit
    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:

    ·