This project focuses on improving banking customer support through AI-powered intent classification and response generation. We utilized TF-IDF, FastText, GloVe, and Sentence Transformer embeddings combined with Cosine Similarity, Euclidean Distance, and Manhattan Distance to accurately understand user queries. The system integrates banking APIs for transactional tasks and employs large language models (LLMs) for handling detailed inquiries. To further enhance performance, we expanded the dataset by adding more banking-related intents, ensuring a more comprehensive and effective user experience.
Tools: VS Code, NLTK, Python, Qdrant, SQLite, Llama 3, Excel, Google Colab
Department: Department of Mathematics
Project Poster
