3 ms·
Yeah, the translation hasn't quite been massaged yet. The next steps are; 1) Complete copying the dictionary in 2) Add more variations of words e.g. hi, hello
by thomasfromcdnjs 6y ago
Yeah, the translation hasn't quite been massaged yet. The next steps are;
1) Complete copying the dictionary in
2) Add more variations of words e.g. hi, hello, gday
3) Add a probability matcher that looks for % matches of liklihood e.g. house != hosue (0.75)
4) Use the saved grammar types to try infer if the word is a verb before noun etc
5) Use a real ML/NLP system to really try understand the language
rinse and repeat those steps for optimizations. A few developers have already reached out to contribute so hopefully get this up to scratch asap.