but they are slow and not as effective. There's no magic to it, but there is a serious problem with structured data sets. To train an algorithm to solve a real-world problem, you have to feed it data about hundreds of thousands or millions of different, but similar, things. For example, to create a program for caring for cows on a farm, you will have to upload all their photos and recordings of mooing to the neural network.
And if you are going to filter spam, try not to forget about Nigerian letters . israel email list All this wealth must first be prepared: find all your cows in the photos and outline them with something bright - so that the algorithm understands where their border with the background is. And phrases with the meaning "urgent help to get a million dollars" - mark with a marker or something like that.
Without preliminary preparation - marking, data is of little interest to anyone except you. The computer itself cannot do this, it needs people. WE WILL PROMOTE YOUR BUSINESS Read more Army of free helpers When someone wants to solve a captcha , they are asked to find a dozen or so intersections, traffic lights, or shop windows in photographs.
Simple algorithms are also suitable
-
- Posts: 506
- Joined: Sat Dec 28, 2024 3:24 am