LlamaIndex Newsletter 2025-01-21

Hi there, Llama Fans! 🦙

Welcome to this week's edition of the LlamaIndex newsletter! We're thrilled to share exciting updates, including insights into the AutoRAG framework and strategies for enhancing knowledge graph applications.

🤩 The Highlights:

• In Case You Missed It: week before last we released a New Multilingual Visual Embedding Model: We open-sourced our visual embedding model and training set on Huggingface! Key features include:
  • Trained on 5 languages (IT, ES, EN, FR, DE)
  • 70% fewer tokens = 3x faster inference for the English-only model
  • Matryoshka Representation Learning for flexible dimension reduction
  • Full training set is open-source
  • Learn more here and check out the model integration.
• Introducing AutoRAG Framework: AutoRAG is a framework for optimizing your RAG pipelines. Key findings include:
  • Hybrid retrieval methods often outperform pure vector or BM25 approaches.
  • Query expansion isn't always beneficial—context matters.
  • Some rerankers may perform worse than no reranking—test before implementing.
  • Read the full paper here.
• Agentic Strategies for Knowledge Graphs: Dramatically improve the accuracy of your knowledge graph applications by applying agentic strategies with LlamaIndex workflows! Key insights include:
  • Implement agentic strategies for text2cypher using LlamaIndex Workflows.
  • Explore multi-step approaches with retry and self-correction mechanisms.
  • Check out the full guest post here.

🗺️ LlamaCloud & LlamaParse:

• Building RAG Applications: Learn how to build a RAG application using LlamaParse, LlamaCloud, and AWS Bedrock. This guide covers:
  • Efficient document parsing with LlamaParse.
  • Managing indices on LlamaCloud.
  • Integrating AWS Bedrock's embedding models and LLMs.
  • Check it out here.

✨ Framework:

• Agentic Strategies Implementation: Discover how to implement agentic strategies for text2cypher using LlamaIndex workflows. This comprehensive post covers:
  • Error checking, retries, and correction mechanisms.
  • Benefits of iterative planning for complex queries.
  • Read more here.
• AutoRAG Framework Insights: Explore the AutoRAG framework for picking optimal configurations for your RAG pipelines. Key findings include:
  • Evaluation techniques and data sets used.
  • Read the full paper here.

✍️ Community:

• AI Tour Planner Agent: Build a travel planner agent from scratch with @clusteredbytes using LlamaIndex workflows and SerpAPI. Check it out here.
• Women in AI RAG Hackathon: Join the Women in AI RAG Hackathon in Palo Alto to explore RAG using open-source vector databases. Register now.
• Webinar Recap with Memgraph: Learn how @memgraphdb and LlamaIndex work together to build agentic graph applications. Read the recap or watch the recording here.
• Neomagus LLM x Law Hackathon Winner: Discover how Neomagus won the hackathon with a solution for verifying AI-generated legal information. Read the full post here.

Thank you for being a part of our community! Stay tuned for more updates, and don’t hesitate to reach out with your questions or feedback.

Happy Llama-ing! 🦙