Engineering & simulationcontrastive learning model77% success rate for single pairs; 88% with multiple samples
An undergraduate student trained a contrastive learning model on a public database of 60,000 fingerprints to determine if different fingers from the same person share identifiable similarities, achieving a 77% success rate for single pairs and up to 88% with multiple samples.
“the, uh, artificial intelligence model that they decided to pursue here is the contrastive learning uh, model. … What they ended up realizing, um, as this thing was trained was, I think with a 77% certainty or positive rate, it could tell us that, you know, different fingers from the same person belonged to the same person…”
Speaker 1 · Hear it at 7:45