Inspiration
Connections Bot was inspired by the NYT connections game, aiming to help players group words based on their meanings. We used Word2Vec to understand the relationships between words and create intelligent suggestions.
What it does
The bot groups words by their semantic similarity using Word2Vec, helping players find connections in a word puzzle. It tracks previous guesses and adjusts its suggestions accordingly, even using "one-away" guesses to improve results.
How we built it
We used Python and Gensim’s Word2Vec model to calculate word similarities. The bot evaluates all possible word combinations, ranks them, and suggests the most relevant groups. It also tracks guesses and adapts to improve future suggestions.
Challenges we ran into
We faced issues with words not found in the Word2Vec model, and the algorithm needed optimization for efficiency, especially with larger datasets. Balancing the influence of previous guesses was also tricky.
Accomplishments that we're proud of
We successfully created a system that groups words based on their meanings, adapts to user input, and can scale to larger puzzles.
What we learned
We learned how to work with Word2Vec for semantic analysis, optimize algorithms for efficiency, and integrate game mechanics with an adaptive bot.
What's next for Connections Bot
We plan to improve user feedback integration, expand game features, and enhance the bot’s learning capabilities for more personalized suggestions.
Log in or sign up for Devpost to join the conversation.