Topic on UpTrust
data science
A live feed of posts and discussions about data science, ranked by who the most trusted people trust, not by likes or follower counts.
Log in and start voting to get personalized rankings.
6 posts1 trusted voiceUpdates live
Daily Alchemy: a question to think on together
“When a model gives you a confident answer that turns out to be wrong, do you trust the tool less or trust yourself more?”
Trust ranks through the site stewards' curated lens.
You're correct that the trust tracks different groups of people rating expertise differently. there are two things we haven't fully released yet that address trust bubble fragmentation: an algorithmic way of identifying the people you most trust who have a different view from you... - dave...
I’ve had some chats with someone who really didn’t like the believability scores and had feelings about how it all went down, and the impression I got was that it was coming from a more teal place. I might be getting the colour scheme wrong there though.... The past two weeks at UpTrust (Sep 9)
Lots of data science-y stuff, improving the core algorithms, showcase sites https://science.uptrusthq.com/ and https://tpot.uptrusthq.com/, plus... Writing and Reading Long posts get a full page at /posts/new instead of a cramped box.... - rnicholson...
My name is Robert. I'm on the faculty at San Jose State University, where I teach classes in computer science, computer engineering, and data science.... I dig it, good idea! This is the kind of thing LLMs are usually good at, and something we’ll have the data for. we’ll still need to prompt it as to what kinda of frameworks to use or else it’ll default to telling you a particular developmental levels values, but that’s not so bad...