Natural Language Processing in Apple Ecosystem

Thanks for the interesting video.

Best video I've seen on Swift NLP. Is it possible to set up custom tokenizers in swift? For example I'm trying to tokenize the city named Mariupol in Chinese, which should be 馬利烏波爾 . But when tokenized in swift it comes out as 1: 馬利, 2: 烏, 3: 波爾 Is there a way to let swift know I want the token as 馬利烏波爾 . Thanks again for your informative video!

Great overview! I’m building my final year project and you’re tutorial is helping me a lot

Can you share your knowledge regarding NLP resources? Websters Unabridged dictionary contains approximately 500 000 defined words. How does Apple's development ecosystem store these? To determine parts of speech (verb, noun, etc) I imagine these must be stored in a file? To determine their proximity to other words (couch, settee, chair) these are also stored? Where are these NLP resources stored? If I include NLP in my app, are these resources also copied into my app's bundle when loaded onto an iPhone?

How do these NLP resources get updated with new definitions, words, and refinements?

Is there a similar set of files for every language in the NLP ecosystems (Polskie, Italiano, etc) ??

Appreciate any insight you can provide regarding where these NLP resources are stored, and how they are used.

Excellent summary! You gave a great overview of NLP where the words in the examples are bonafide words. In business writing, this is seldom the case. Business writing is plagued with abbreviations and acronyms. One of the most odious chores for technical editors is creating an appendix listing and defining every abbreviation in a document.

I suspect NLP can help identify abbreviations? A common pattern would be ALL CAPS. Examples: B2B, SMTP, CSS.

One approach might be to determine if the word is spelled correctly? SMTP is not a word in English, same with CSS.

What approach could you recommend using NLP to identify abbreviations and acronyms in a large document?