Prior Labs and the end of dumb data

The founder of Prior Labs, Frank Hutter, changed the field of data science once with his work on AutoML, which automated the end-to-end process of building machine learning models
He and his team are back to do it again with TabPFN, the most performant tabular foundation model available. In their recent Nature paper, TabPFN was shown to outperform the accuracy of state-of-the-art models in over 96% of use cases on small tabular data. It requires 50% of the data to reach the same level of accuracy as the next best model and only takes 2.8 seconds to deliver better performance than the best existing models in 4+ hours. It really is the era of opensource-AI dominance. And it continues to astound me just how effective end-to-end deep learning is in almost any field it is applies too.
TabPFN has many advantages, but is still remarkably easy to use and can be applied to any dataset quickly with just a few lines of code. Since it's release just a few weeks ago, we've seen TabPFN's opensource solution outperform other ML techniques in areas as diverse as clinical trial data, commodity trading and retail forecasting.
Going forward, as the size of datasets that TabPFN can handle scales, we expect to see this methodology having an even more transformative effect on datascience and as a result how enterprises and scientists get value from their data. We are very close to being able to quickly understand the relationship of data across data types, columns and tables allowing even the most novice analysts to interrogate data in a way that would take dozens of experts and huge resources today.
If you want to try it out, they've released their first model under an Apache license and have an API available at www.priorlabs.ai. Try it out or send it to any data scientists you know, they'll thank you later.
Congratulations to Frank, Noah and Sauraj on an incredible technical achievement and launch! We're proud to be working with them alongside XTX Markets, as well as the Hector Foundation, Atlantic Labs, Galion.exe, Thomas Wolf, Peter Sarlin, Guy Podjarny, Ed Grefenstette, Robin Rombach, Chris Lynch, Ashutosh Kulkarni and many more great angels.You can learn more in this great Fortune piece here: https://lnkd.in/eUBNXe8g