02 Dec 2025
49m

Technical advances in document understanding

Podcast cover

Practical AI

In this episode of the Practical AI podcast, co-hosts Daniel Whitenack and Chris Benson discuss the evolution and practical applications of AI in document processing. They explore various modeling techniques, including OCR, document structure models like Dockling, language vision models, and DeepSeek OCR, highlighting their strengths, limitations, and use cases, particularly in enhancing RAG systems. The conversation emphasizes the importance of preserving document structure and context for improved AI performance, contrasting traditional methods with newer, more innovative approaches that address resolution and data representation challenges.

Outlines

Part 1: Introduction and Context

Part 2: OCR and Document Structure Models

Part 3: Language Vision Models and DeepSeek OCR

Part 4: Conclusion and Gratitude

Sign in to continue reading, translating and more.

Open full episode in Podwise