Banking77 intent taxonomy.
Semantic Banking Assistant
A semantic assistant that uses sentence-transformer embeddings and similarity search to understand Banking77 intents and return conversational responses.
Connected an embedding-based intent workflow to a focused Streamlit interface, turning an NLP experiment into an end-to-end demonstration with thresholded semantic retrieval.
01 / PROBLEM
What the system is trying to solve.
Intent systems based only on exact keywords are brittle. The project explores whether sentence embeddings can provide a compact semantic layer for routing natural banking questions into a fixed intent taxonomy.
02 / SYSTEM
How it works.
- 01
Embed Banking77 examples with a sentence-transformer model.
- 02
Compare incoming user queries against intent representations with cosine similarity.
- 03
Apply a confidence threshold and return the strongest semantic matches.
- 04
Cache model resources and keep lightweight session history in Streamlit.
- 05
Expose top-k alternatives so uncertain intent matches are inspectable instead of hidden.
03 / ENGINEERING
Decisions that shaped the project.
Sentence-transformers with all-mpnet-base-v2 embeddings.
Cosine-similarity retrieval with thresholding.
Top-k intent inspection and conversational session state.
Small, deployable Streamlit interface.
04 / EVIDENCE
A few concrete signals.
Banking77
Datasetall-mpnet-base-v2
Embedding modelCosine similarity
RetrievalStreamlit
Interface05 / LINKS