Ria Kalluri

@riakall

phd studying AI/ML at stanford, thinking about the concepts inside machine learning models + dreaming up more radical ai 🌱

Stanford, CA
Vrijeme pridruživanja: prosinac 2018.

Tweetovi

Blokirali ste korisnika/cu @riakall

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

  1. Prikvačeni tweet
    22. lis 2019.

    The top 5 readings that are helping me dream up radical AI right now (a thread) 1/5 A People's Guide to AI by + a guide to AI for anyone, that's "as focussed on equity as it is on tech". It's just so good.

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    30. sij

    Believing that only computation can perceive, process, &understand the physical world, implies that “ can know people better than they can know themselves". We ascribe it a "logic of revelation" instead of mere “approximation” ht

    Poništi
  3. proslijedio/la je Tweet
    27. sij
    , , i još njih 6
    Poništi
  4. proslijedio/la je Tweet
    15. pro 2019.

    Please stop describing AI safety as "long term, technical" safety. Ethics, fairness and AI for Social Good problems are also very much playing the long game & in many cases involve confronting highly technical fundamental ML challenges, requiring theoretical innovation to solve.

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    10. pro 2019.

    Friends, a website! and I hereby announce the new website for our forthcoming book, Data Feminism. And most notable about the website is the PRE-ORDER link: cc:

    Poništi
  6. proslijedio/la je Tweet
    11. pro 2019.

    📢📢📢 The Feminist Data Manifest-No: 📢📢📢 Read the short version, read the long version, read & share them both! By:

    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    11. pro 2019.

    💭❓[Thread:] How can the field of computing contribute to social change, without tacitly assuming that tomorrow's social structures will basically resemble the past? ⚡️

    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    9. pro 2019.

    "Is this ML model doing good?" is the wrong question to be asking. We need to ask: "how is this ML model shifting power?"

    "Is this ML model doing good?" crossed out.

"how is this ML model shifting power?"
    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    10. pro 2019.

    More takeaways from this amazing work on algorithmic injustices by 1) Centre society's most vulnerable rather than the technology. 2) Automated decision-making systems create and sustain a certain social order. 3) Prioritise understanding rather than predictions.

    Poništi
  10. proslijedio/la je Tweet
    9. pro 2019.

    It is increasingly difficult to find an aspect of lived experience that is not datafied, measured, quantified, and analyzed.

    It is increasingly difficult to find an aspect of lived experience that is not datafied, measured, quantified, and analyzed.
    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    9. pro 2019.

    As a field, we are overly focused on prediction. We need to prioritize understanding over prediction. It's not about the "right answers" but asking meaningful questions

    We need to prioritize understanding over prediction. 
Ask "why do we find the patterns that we do?" Instead of using perceived patterns as inputs towards predictive systems.
It's not about the "right answers" but asking meaningful questions.
    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    8. pro 2019.

    Super excited to be in conversation at the workshop about and in the context of and — looking forward to de-smokescreening the discourse!

    Poništi
  13. 22. lis 2019.

    Would love to hear what else/who else I should definitely be listening to!

    Prikaži ovu nit
    Poništi
  14. 22. lis 2019.

    5/5 Counting the Countless post by a clear & cutting critique of data science as fundamentally reductive & coercive. ends with a brief nod to the wish for "a radical data science" that values context, autonomous control & many ways of being

    Prikaži ovu nit
    Poništi
  15. 22. lis 2019.

    4/5 The Moral Character of Cryptographic Work by Phillip Rogaway not about AI. discusses researchers’ relationship to power, how centralized data & cryptography research rearrange power, and which cryptography research may do so in a just way

    Prikaži ovu nit
    Poništi
  16. 22. lis 2019.

    3/5 Design Justice, A.I., and Escape from the Matrix of Domination by a concrete story of values encoded in AI, which segues into an intro to the matrix of domination, and a case for for AI that makes space for many ways of being

    Prikaži ovu nit
    Poništi
  17. 22. lis 2019.

    2/5 Data Feminism, ch 1 by + Intro principles of feminist data science including "the view from nowhere [feigned objectivity] is always a view from somewhere [often power]" that should be stated. Again: table near the end is gold!

    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    26. lip 2019.

    Children use the mutual exclusivity (ME) bias to learn new words, while standard neural nets show the opposite bias, hindering learning in common scenarios. New preprint from introduces ME as a challenge for neural networks

    Poništi
  19. proslijedio/la je Tweet
    26. lip 2019.

    Exited to share our new paper: 'Monte Carlo Gradient Estimation in Machine Learning', with @elaClaudia . It reviews of all the things we know about computing gradients of probabilistic functions. 🐾Thread👇🏾

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    8. tra 2019.

    Style transfer without a specific style image. I minimize only the content loss from neural style transfer and directly optimize brushstrokes via backpropagation. In this case, "style" is an intrinsic property defined entirely by the artistic medium.

    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:

    ·