Automating Paperwork
Warning: This book was written before LLMs exploded onto the scene in a production-ready way. I no longer consider the advice within current with respect to recent developments.
- Ted, Jan 2023
A book for business and engineering leaders seeking to understand how document automation will impact their business and how to win with this new technology. Strategic, practical, and actionable.
- Ted, Jan 2023
A book for business and engineering leaders seeking to understand how document automation will impact their business and how to win with this new technology. Strategic, practical, and actionable.
👉 Getting started
📝 Natural Language Understanding in the Enterprise
What is Natural Language Understanding?
How to find NLU use cases likely to succeed?
🔎 Vetting a Document Extraction Use Case
The most important work happens before computers are involved
Defining business problems by their inputs
Vetting the documents available for automation
⚙️ Extraction Pipelines
Quality Assurance and Feedback Loops
🧠 Models from a 10,000 Foot View
Models from a 10,000 Foot View
Learning representations and solutions
Parameterization and its structure
Who will train and update my model?
What is my model capable of learning?
Should I pick the newest model?
📈 Trends and Predictions
Lower training data requirements
Integrated development environments
Layout and Image Understanding
Supporting complex outputs and domain constraints
Talking to models instead of invoking them
OCR will approach human levels