Paradigm Challenge

Paradigm Challenge

1618 papers · Page 17 of 17

Distilled datasets often fail to beat random image selection once the soft label trick is removed from the equation.

AI & ML arxiv | Apr 24

AI models show much higher levels of bias when generating complex machine learning code than they do when writing simple "if-then" statements.

AI & ML arxiv | Apr 24

A new medical diagnostic tool reveals that AI models still rely on scientifically debunked racial myths to make patient health predictions.

AI & ML arxiv | Apr 24

Intelligence is actually just the process of extreme data compression, and a "V-shaped" pattern in a model's layers proves it.

AI & ML ssrn | Apr 24

There is no single "lying center" in an AI. a model's tendency to hallucinate is controlled by different neurons depending on the subject.

AI & ML arxiv | Apr 24

57 percent of users have zero or negative correlation with the global leaderboards used to rank AI models.

AI & ML arxiv | Apr 24

Prompt engineering cannot recover information that a human user never put into the text.

AI & ML arxiv | Apr 24

Sensitive images of secret computer chips can be reconstructed from encrypted updates even when the data never leaves the original server.

AI & ML arxiv | Apr 24

A new attack method called ProjRes identifies whether a specific person's data was used to train a model with nearly 100% accuracy.

AI & ML arxiv | Apr 24

Multimodal AI models are often "functionally blind," guessing what is in an image based on the text instead of actually looking at it.

AI & ML arxiv | Apr 24

Knowledge graph models are 25 percent dumber at remembering things than previous benchmarks suggested.

AI & ML arxiv | Apr 24

Complex architectural upgrades for small AI models are completely unnecessary when simple, well-chosen examples work just as well.

AI & ML arxiv | Apr 24

The most "active" parts of an AI's brain are almost entirely unrelated to the actual decisions the AI makes.

AI & ML arxiv | Apr 24

A widely accepted mathematical assumption used to speed up AI decision-making algorithms has just been proven wrong.

AI & ML arxiv | Apr 24

AI agents can now "realize" when they are thinking about a problem the wrong way and restructure their entire mental model on the fly.

AI & ML arxiv | Apr 24

Smaller, bio-inspired controllers actually outperform massive AI networks for the task of moving a robot's legs.

AI & ML arxiv | Apr 24

A 1965 theorem about simple networks fails completely when you apply it to the complex, multi-way relationships of a hypergraph.

AI & ML arxiv | Apr 24

A superintelligent system that constantly improves itself will eventually destroy its own identity through a mathematical loop of self-modification.

AI & ML arxiv | Apr 24