• 0 Posts
  • 23 Comments
Joined 1 year ago
cake
Cake day: June 17th, 2023

help-circle










  • Zeth0s@lemmy.worldto4chan@lemmy.worldAnon on credit scores
    link
    fedilink
    arrow-up
    6
    arrow-down
    1
    ·
    edit-2
    1 year ago

    It was just to give an idea that what OP mentioned is already an established thing, fairer than alternatives.

    Most of the time trivial linear logistic regression is used in this context. Nowadays decision tree ensambles are pretty heavily used, which are ML. Simply they perform better with fewer data than neural networks on structured tabular data.

    What you refer to as AI is probably methods based on deep learning. The truth is that they work exactly as any other algorithm that you are referring to. They are used for regression and classification, same way as a standard linear regression. The difference is that the models are non linear, and their complexity is so that a lot of data are needed to train them.

    But conceptually one can absolutely create a credit score with deep neural networks. It is just an overkill, for performances that are likely worst than a random forest on relatively small training datasets

    Neural networks-based methods are indeed used in fraud detection




  • Zeth0s@lemmy.worldto4chan@lemmy.worldAnon on credit scores
    link
    fedilink
    arrow-up
    11
    arrow-down
    1
    ·
    edit-2
    1 year ago

    You know there are people in bank and credit institutions that have been doing this for centuries? Probably millennia… EU explicitly requires that some of this is done by what you call AI (i.e. mathematical models) because they are fairer than humans and safer for customers and society

    Check basel III for an intro on the topic





  • The real challenge is “how do users can judge what is a fake news?”. In a similar situation it is an extremely difficult task even for newspapers with journalists on the field. See what’s happening with the blame-shifting on the bombing of Gaza’s hospital.

    Even guardian and bbc have trouble understanding where is the truth.

    A solution could be filtering the sources (for instance, no unknown blogs, or the sun and fox News, only reputable sources such as guardian and bbc). But important real news might be missed in this case, that are direct testimony of journalists on the field. And supposedly reputable sources such as wsj or similar are also known to have shared fake news, particularly when it comes to this conflict. And also reputable sources are biases.

    It is an extremely difficult topic. No one has a definitive answer unfortunately.

    I would be in favor of filtering at least the widely known sources of fake news (shady blogs, all Murdock’s media and so on)

    Edit. An adjective to clarify