Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Week 10 #1

Open
twood02 opened this issue Nov 8, 2016 · 2 comments
Open

Week 10 #1

twood02 opened this issue Nov 8, 2016 · 2 comments

Comments

@twood02
Copy link
Contributor

twood02 commented Nov 8, 2016

No description provided.

@twood02
Copy link
Contributor Author

twood02 commented Nov 8, 2016

Demo: gathering data from site and sample output from template

Next steps: construct templates. How to combine templates? What would you train? How to use sentence structure data? How to generate new examples instead of just use classify them?

@sutardji
Copy link
Contributor

Switched over from generating updates from games in real-time to focus on aggregating data and relating plays to overall stats.

Effort: More NLP, less technical.

Reading:

Problems:

  • Templates seem to be the best option for relating data. The English language is taught through grammatical templates in order to make sense of words, but language becomes more flexible the more it is understood. Sentence diversity seems to direct to providing more templates.
  • Thinking about having program generate sentences from a list of grammatical "don'ts", rather than "do's". Learning from negatives seems easier and more accommodating to the flexibility of sentence structure.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants