https://github.com/rossgoodwin/lexiconjure

Twitter bot generating invented words and definitions using RNN + genetic algorithm
https://github.com/rossgoodwin/lexiconjure

Last synced: 1 day ago
JSON representation

Twitter bot generating invented words and definitions using RNN + genetic algorithm

• Host: GitHub
• URL: https://github.com/rossgoodwin/lexiconjure
• Owner: rossgoodwin
• License: other
• Created: 2016-02-17T18:45:30.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2016-02-21T19:05:35.000Z (over 8 years ago)
• Last Synced: 2024-08-02T05:11:54.042Z (3 months ago)
• Language: Python
• Homepage: https://twitter.com/lexiconjure
• Size: 475 KB
• Stars: 131
• Watchers: 7
• Forks: 9
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# lexiconjure

[@lexiconjure](https://twitter.com/lexiconjure) is a twitter bot that uses machine learning to define invented words, posting truncated definitions on Twitter and complete ones [on Tumblr](http://lexiconjure.tumblr.com).

It uses an [LSTM](https://en.wikipedia.org/wiki/Long_short-term_memory) neural network trained on the [Oxford English Dictionary](http://www.oed.com/). When you tweet a word (real or made up) @ the bot, you're seeding that generator.

By smashing together random sequences of characters with a genetic algorithm until something that looks (sort of) like a real word comes out, the bot invents its own words to define every 90 minutes.

### Credits

* Bot created and trained by [Ross Goodwin](http://rossgoodwin.com)

* LSTM is [Char-RNN](https://github.com/karpathy/char-rnn) by Andrej Karpathy

* Linguistic-taste integrity moderated by badwords.json via [Darius Kazemi](https://github.com/dariusk/wordfilter)

* Model to be released under creative commons when training is complete

* Profile image: [Book by Julien Deveaux from the Noun Project](https://thenounproject.com/Julihan)