TrivialAugment: The next evolution in Data Augmentation [Breakdowns]
Devanshartificialintelligencemadesimple.substack.com![Thumbnail of TrivialAugment: The next evolution in Data Augmentation [Breakdowns]](https://substackcdn.com/image/fetch/w_1200,h_600,c_fill,f_jpg,q_auto:good,fl_progressive:steep,g_auto/https%253A%252F%252Fsubstack-post-media.s3.amazonaws.com%252Fpublic%252Fimages%252Fd743b381-7724-4cff-80d9-ba3bc905f4b7_700x483.png)
TrivialAugment: The next evolution in Data Augmentation [Breakdowns]
The case for Reality-centric AI // van der Schaar Lab
vanderschaar-lab.com
Incredible paper. One weird trick to recursive self-improvement:
- a model iteratively labels its own train data and learns from progressively harder examples. Gotta
1) generate problems of appropriate hardness
2) be able to filter our negative examples using a cheap verifier. https://t.co/nkYqGgNM7q