A degree of image identification at sub-human scales could be possible with more advanced clusters

TL;DR

This research investigates whether self-supervised learning with increased data volume and image quality can achieve human-level visual comprehension at sub-human scales, demonstrating promising results with vision transformers.

Contribution

It introduces a scaling experiment that combines data volume and image resolution increases to reach human-level detection performance without external funding.

Findings

Scaling data and image quality improves detection accuracy.

Vision transformers achieve human-like performance at small scales.

Sub-human size detection becomes feasible with advanced clustering.

Abstract

The purpose of the research is to determine if currently available self-supervised learning techniques can accomplish human level comprehension of visual images using the same degree and amount of sensory input that people acquire from. Initial research on this topic solely considered data volume scaling. Here, we scale both the volume of data and the quality of the image. This scaling experiment is a self-supervised learning method that may be done without any outside financing. We find that scaling up data volume and picture resolution at the same time enables human-level item detection performance at sub-human sizes.We run a scaling experiment with vision transformers trained on up to 200000 images up to 256 ppi.