Semantic Search Using Embeddings

A lightweight NLP project that demonstrates semantic search using Hugging Face sentence embeddings and cosine similarity.

The application converts both documents and user queries into vector embeddings and retrieves the most semantically relevant document based on contextual similarity.

---

Tech Stack

• Python
• LangChain HuggingFace
• Sentence Transformers
• Scikit-learn

---

Workflow

• Generate embeddings for documents
• Generate embedding for user query
• Compute cosine similarity scores
• Retrieve the closest matching document

---

Project Structure

semantic-search-using-embeddings/
│
├── app.py
├── requirements.txt
├── README.md
└── .gitignore

---

Installation

git clone https://github.com/your-username/semantic-search-using-embeddings.git

cd semantic-search-using-embeddings

pip install -r requirements.txt

python app.py

---

Sample Query

query = "Who is the best bowler?"

Retrieved Result

Jasprit Bumrah is widely recognized for his unique bowling action and deadly yorkers in death overs.

---

Concepts Used

• Sentence Embeddings
• Semantic Similarity
• Cosine Similarity
• Vector Retrieval

---

Future Scope

• Vector database integration
• Streamlit interface
• PDF/document retrieval
• RAG-based search pipeline

---

Vrundali Rahangdale