Text classification with vector embeddings — and no ML model

[curt.kennedy]

Yes, and it works with images using image embedding models too.

And when you increase the SNR by weighting with correlation values across multiple hits … LOOK OUT!

It is one of those hidden gems.

This works for linearly related labels using embeddings. And gives a confidence factor for the label (sigma).

What is a proper way to combine multiple cosine similarities?

For simplicity, you could take the sum of all the vectors, then normalize it back to the unit hypersphere. This would essentially represent a new embedding vector that is the average embedding. Let S = \sum v_i, for vectors v_i. Then V = \frac{S}{||S||} To figure out what this new vector represents, correlate it with all your previous labeled embedding vectors. Then take the average or top label as your answer. If your categories are linearly related, like (0) Awesome, (1) Good, (2) OK, (3…