Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
7 views12 pages

Understanding Retrieval-Augmented Generation (RAG)

Retrieval-Augmented Generation (RAG) is a hybrid AI framework that combines retrieval of relevant information and generation of coherent answers, enhancing traditional AI models by grounding responses in specific documents. It involves data preparation, query processing, retrieval, generation, and delivery of answers, making it applicable in various fields such as customer support and research. While RAG offers accurate and scalable answers, its effectiveness depends on the quality of the knowledge base and can incur computational costs.

Uploaded by

Rohit Srivastava
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
7 views12 pages

Understanding Retrieval-Augmented Generation (RAG)

Retrieval-Augmented Generation (RAG) is a hybrid AI framework that combines retrieval of relevant information and generation of coherent answers, enhancing traditional AI models by grounding responses in specific documents. It involves data preparation, query processing, retrieval, generation, and delivery of answers, making it applicable in various fields such as customer support and research. While RAG offers accurate and scalable answers, its effectiveness depends on the quality of the knowledge base and can incur computational costs.

Uploaded by

Rohit Srivastava
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

Understanding

Retrieval-Augmented
Generation (RAG)
What is RAG? A Real-Life
Analogy

► Imagine a super-smart librarian in a vast library.


► Someone asks: “What are the benefits of a
Mediterranean diet?”
► Librarian searches the catalog, pulls relevant books,
and summarizes key points.
► RAG does this digitally: searches data and
generates
answers.
What is Retrieval-Augmented Generation
(RAG)?

► A hybrid AI framework combining:


► Retrieval: Finds relevant info from a dataset.
► Generation: Crafts coherent answers using AI.
► Used in chatbots, search tools, and research
assistants.
► Ensures answers are accurate and up-to-date.
Why Do We Need RAG?

► Traditional AI models have limitations:


► Limited Knowledge: Fixed at training time, can be
outdated.
► Hallucination: May generate incorrect info.
► Lack of Specificity: Struggles with domain-specific
data.
► RAG solves this by grounding answers in specific
documents.
How Does RAG Work?

1. Data Preparation: Build a knowledge base


(documents, PDFs, etc.).
2. Query Processing: User asks a question (e.g., “How
many vacation days?”).
3. Retrieval: Find relevant document chunks.
4. Generation: AI crafts a coherent answer.
5. Delivery: Return answer with source references.
Step 1: Data Preparation

► Collect Documents: E.g., employee handbook,


research papers.
► Chunking: Split into small pieces (100500 words).
► Embeddings: Convert chunks to numerical vectors
using models like BERT.
► Indexing: Store embeddings in a vector database
(e.g., FAISS, Pinecone).
Step 23: Query Processing and
Retrieval

► Query Embedding: Convert user question to a vector.


► Similarity Search: Compare query to document
embeddings using cosine similarity.
► Retrieve Top-K Chunks: E.g., top 5 most
relevant document pieces.
► Example: Query “How many vacation days?” retrieves
“20 days per year” chunk.
Step 45: Generation and Delivery

► Augmentation: Combine query and retrieved chunks


into a prompt.
► Generation: Use a language model (e.g., GPT,
LLaMA) to create an answer.
► Example Answer: “Employees get 20 vacation days per
year.”
► Delivery: Return answer with source (e.g.,
“Employee Handbook, Section 4.2”).
Benefits and
Limitations

Benefits: Limitations:
► Accurate, ► Depends on quality
fact-based of knowledge
answers. base.
► Flexible for any domain. ► Retrieval errors can
► Scalable with large affect answers.
datasets. ► Computational cost for
► Transparent with large datasets.
source references.
Real-World Applications

► Customer Support: Chatbots answering from


manuals or FAQs.
► Enterprise Search: Querying internal policies or
specs.
► Research: Summarizing papers with citations.
► Legal/Healthcare: Retrieving case law or medical
literature.
References

► Lewis, P., et al. (2020). “Retrieval-Augmented Generation


for Knowledge-Intensive NLP Tasks.”
https://arxiv.org/abs/2005.11401
► Gao, Y. (2023). “What is Retrieval-Augmented
Generation (RAG)?” Towards Data Science.
https://towardsdatascience.com/
what-is-retrieval-augmented-generation-rag
► Pinecone (2024). “RAG: The Definitive Guide to
Retrieval-Augmented Generation.”
https://www.pinecone.
io/learn/retrieval-augmented-generation/
Questions
?

Thank you! Any


questions?

You might also like