4 ms·
You might try vector quantization (instead of PCA) if you just need your 768 features to be smaller. ML features tend to be robust to some perturbation.
by sdenton4 7y ago
You might try vector quantization (instead of PCA) if you just need your 768 features to be smaller. ML features tend to be robust to some perturbation.
- binarymax 7y agoWell it’s one problem or another. If you compress too much you lose the value, and if you leave it too large you have the size problem. Inverted indices are very efficient. How much of that can you give up at what trade off? If I’m only going to be better for 10% of queries, is that a cost effective solution? What if I spend the same amount of time tuning a traditional engine a bit more and get better accuracy for 5% of queries? Tradoffs rule the world of practical search implementations.