Skip to content

Explaining NLP through Machine Learning and a visually appealing experience.

Notifications You must be signed in to change notification settings

pelinzon/ExploringNaturalLanguage

Repository files navigation

Exploring Natural Language

Thumbnail

This project is a Swift Playground that explains core machine learning and Natural Language Processing (NLP) concepts and showcases Apple's new Natural Language framework through a visually appealing experience. Some of the technologies used were CoreML, UIKit, PlaygroundBook, CreateML and the framework itself, along with three different publicly available datasets.

A demo of the project running on iPad is available here on Youtube.

I used this project as a submission to the WWDC19 Scholarships program and it was accepted. As a reward, I got to attend WWDC 2019 with my travel expenses covered by Apple and meet the engineers who built the frameworks we use daily.

Screenshot

Pre-Compiled Instructions

  1. Download the Exploring Natural Language.playgroundbook file;
  2. AirDrop the file to your iPad;
  3. Open with Swift Playgrounds and have fun!

Compiling Instructions

  1. Open the project using Xcode;
  2. Select the PlaygroundBook target and build it;
  3. At the very last folder, named Products, you should see a new .playgroundbook file;
  4. Right-click it, and click on Show on Finder;
  5. AirDrop the file to your iPad;
  6. Open with Swift Playgrounds and have fun!

What I’ve Learned

  • Important Machine Learning & NLP concepts
  • How to use Natural Language framework
  • How to use CreateML to train text recognition models
  • How to create engaging learning experiences using Swift Playgrounds

Requirements

  • Xcode 10 for compiling (use Swift 4.2 toolchain if needed)
  • iPad running iOS 12+
  • Swift Playgrounds 2.0+

Project Information

  • Created: 04/2019
  • Updated: 06/2019

About

Explaining NLP through Machine Learning and a visually appealing experience.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published