Automate your document pipeline with intelligent document processing
summary
Step-by-step demo builds a serverless pipeline that extracts text with Amazon Textract, runs a custom entity recognizer using Amazon Comprehend (trained on two labels: genuine and fake), routes predictions to Amazon Augmented AI (A2I) for human correction, and retrains the recognizer automatically. Implementation uses S3, Lambda and CloudFormation, processes a PNG loan application end-to-end, shows Comprehend jobs complete and A2I JSON outputs appear, and a scheduled Lambda triggers retraining — the pipeline works as demonstrated.
AI-written summary · about
chapters (9)
- 0:00Introduction and challenges
- 1:34Why use machine learning?
- 7:17AWS AI services overview
- 9:33Architecture and APIs
- 13:57Setup and prerequisites
- 18:56Configuring human review
- 23:05Deploying the pipeline
- 28:42Review and model retraining
- 32:53Resources and next steps
related talks
30mIntelligent document processing: Extract data & insights from documents
26mEmpower your business use cases
42mExtract information accurately from documents using computer vision
32mAutomate data extraction & analysis with intelligent document processing
32mAutomate data extraction & analysis with intelligent document processing
30mIntelligent document processing for the insurance industry
18mUsing AI to automate clinical workflows
43mImplementing AI Document Processing at Scale: Syngenta