Intelligent document processing: Extract data & insights from documents
summary
Pipeline extracts text with Amazon Textract, runs NLP with Amazon Comprehend and routes low‑confidence cases to Amazon A2I; custom image and text classifiers train in SageMaker Ground Truth, convert to TensorFlow Lite and infer in Lambda, orchestrated by Step Functions and surfaced via Amplify/Cognito. System processes 20,000 supplier documents per month, parallelizes per‑page work, and demos a 15‑second verify loop; human labeling (250 workers) closes errors and feeds continuous model retraining.
AI-written summary · about
chapters (7)
- 0:00Introduction and overview
- 4:01Challenges in document processing
- 8:44Extracting data with Amazon Textract
- 14:59Getting insights with Amazon Comprehend
- 21:04Human oversight with Amazon A2I
- 22:27Assent Compliance demo and architecture
- 28:08Future plans and closing
related talks
35mAutomate your document pipeline with intelligent document processing
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
43mImplementing AI Document Processing at Scale: Syngenta
29mSolve common business problems with AI and ML
51mProfundizando en Aprendizaje Automático para Desarrolladores