• In ppoi, I assumed a use case where the rules cannot be verbalized, but sample data can be provided.
  • In some cases, they don’t.
    • There is too much equivalent data that goes into unknown.txt.
    • Create some positives/negatives with miscellaneous rules, and then eyeball them and play the incorrect ones to create teacher data.

This page is auto-translated from /nishio/データかルールか using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.