We had a mixture of talent this time around and decided we could tackle the creative and technical challenge of making a new intuitive way to play a text adventure

What it does

Uses synonyms and natural language processing to figure out actions you can do that are the most similar to the typed sentence.

How we built it

We used the GCP NLP API and the Oxford Dictionaries API to analyze inputted text and compare it to nearest "action" points in feasible action space.

Challenges we ran into

The first challenge we ran into was finding a feasible project for such a diverse team, as we had a CS student, econ student, and English student all in one team. Then, heavy use of internet APIs slows the game down significantly (and took a very long time to figure out). Building a text adventure from scratch involves many bugs, basically none of which were fixed.

Accomplishments that we're proud of

We managed to get a basic proof of concept for a game that intelligently recognizes English sentence inputs and converts them to in game actions.

What we learned

Don't code a text adventure from scratch, and making 12 API calls sure takes a while.

What's next for Silescia

Fix the bugs, add the features we didn't have time to add, and add the rest of the story which we have planned, but not at all implemented. Other potential features in the future would include some ML based intelligent procedural generation of the world, and intelligently intractable NPCs also using the NLP framework.

Share this project: