LemmaLearn

LemmaLearn explores how AI can support language learning by helping users acquire new vocabulary through natural conversation. The experience is designed to make vocabulary practice more adaptive, contextual, and closely aligned with each learner's existing knowledge.

User inputs are lemmatized and embedded using OpenAI, then matched against a Pinecone vector database to surface relevant vocabulary. Words are ranked by semantic similarity, frequency, and the user's existing knowledge, then woven into AI-generated responses to reinforce learning in context and help reduce language fossilization over time.