Machine Learning for Poets: The eMiLy App
summary
A mobile app called eMiLy scrapes about 300 Emily Dickinson poems from Poetry DB, preprocesses them, and scores each line with sentiment so a mood slider can pick “happy” or “sad” poems; the pipeline uses NativeScript + Vue for the front end, Azure cognitive services text sentiment and face APIs for sentiment and selfie mood detection (demo returns an 80% score for “Hope is the thing” and a 100% happy face match), and generates new verses with Markov chains and textgenrnn, sometimes producing plausibly Dickinson‑like lines.
AI-written summary · about
chapters (20)
- 0:00Start
- 1:08Jen Looper Cloud Advocate Lead Microsoft
- 3:58technology and art
- 4:29preservation
- 7:16create new
- 14:51A mobile app to showcase E.D. poetry
- 15:14tools Build the mobile app
- 15:34Microsoft Cognitive Services
- 16:19Azure Machine Learning
- 16:48Scaffold the mobile app
- 19:19Build loadable pages
- 20:00Page 1: Show a daily poem
- 22:01Vocabulary analysis
- 24:51Modern techniques for antique text appreciation
- 25:38REST API call
- 26:27Build a slider
- 28:16Microsoft Face cognitive service
- 29:03match result with poem sentiment
- 31:47Test 1
- 32:35Test 2 text-gen RNN (Max Woolf)
related talks
11mAI-powered JavaScript applications
39mThe art and science of making computers understand sentiment and emotion
39mAI + JavaScript Rocks
43mGetting Sentimental: exploration of iOS sentiment analysis
01h 23mBeyond Sentiment Emotion Mining with Python and machine learning
38mWhat's Love Got To Do With It?...
43mPoetry (almost) From Scratch - Teaching versification to a neural network
11mAmbient Intelligence and Emotionally Intelligent Devices