heiko23.bsky.social
@heiko23.bsky.social
📦 Logistics Coordinator🤖 AI Enthusias🚀 Tech Innovator🌐 Industry 4.0 Advocate🔍 Knowledge Seeker
Reposted
Our work on inductive biases in reinforcement learning is out: tinyurl.com/yfc74e3u

By the fantastic
@noahedrich.bsky.social

Ppl learn faster from slowly changing features; we argue this reflects an inductive bias. Teamwork w @ericschulz.bsky.social & S HallMcMaster
#neuroskyence #compneuro
An inductive bias for slowly changing features in human reinforcement learning
Author summary Learning experiments in the laboratory are often assumed to exist in a vacuum, where participants solve a given task independently of how they learn in more natural circumstances. But h...
journals.plos.org
November 27, 2024 at 8:34 AM
Reposted
Today, November 25th 2024 is the day we officially learned that justice in America, in fact, does NOT matter. #sad
The moral of the story, kids: It’s fine to try to stop the peaceful transfer of power and incite a mob to storm the Capitol and do all you can to overturn an election as long as you run again and win 🇺🇸
November 25, 2024 at 11:13 PM
Reposted
Did you know BlueSky only has 20 employees?

We would like to say a huge THANK YOU to those 20 warriors (who are working 24/7 to keep up with the incredible growth) for giving us a space to share.

We escaped hell to be here.

Thank you, @bsky.app & Jay Graber CEO (not a billionaire)
November 26, 2024 at 5:58 PM
Reposted
TODAY’S EPISODE OF: F*CK AROUND AND FIND OUT!

US farm groups want Trump to spare their workers from deportation.

Gotta ask—how did the farmers vote? It’s almost like actions have consequences…
November 26, 2024 at 1:13 PM
Reposted
This is a feed that aggregates new paper announcements and Arxiv posts. Obviously many are pre-prints and can be hype-ey, so read critically, but it catches some interesting chats and explainer threads. Definitely a work in progress. To have posts show up use #papersky.
November 26, 2024 at 1:21 PM
Reposted
A very interesting paper that examines:

1. Do LLM-based chatbots always produce bullshit?
2. How to reliably detect chatbot bullshit using computational methods?
3. How such detection methods might enable the identification of bullshit in other types of texts?

arxiv.org/pdf/2411.15129
arxiv.org
November 26, 2024 at 7:37 AM
Reposted
You wanted starter packs to be searchable. Our engineers are busy keeping us online, so in the meantime, an independent developer built a new searchable library of starter packs. This is the beauty of building in the open 🦋
November 26, 2024 at 5:11 AM