Colt Sliva

Are these sitelinks? Do only forums with multiple helpful posts on a topic get this feature?

Sifting through data looking for the silver bullet that some random KPI demands

I function well on habit and routine. Here lies my productivity 🪦

@dominos ya’ll baked flies into our pie

No models, no massive compute, and no heavy pretraining. The math is also simple enough it could be done in any program. I look forward to more ingenuity in the cost-effective SEO + ML space! ✌️

The untagged keyword with the shortest distance to a tagged keyword shares the same tag. The code looks complicated, but the results are simple and cheap.

Using the paper, I can start with a training set of keywords that are tagged. Tech and Automotive. Then I have keywords that I want to be tagged for me. For each keyword I want tagged, I add it together with a training keyword and see what the distance is between compressed items

Last week a paper was released showing how gzip could be used for an embedding/compression step. It makes perfect sense when you see how gzip works. aclanthology.org/2023.findin…