Incremental execution for realtime machine learning

--

Summary:
Impira is a system designed to empower users to train their own machine learning models for extracting PDF data, with no code and only a few clicks. But, providing the kind of rapid, iterative retraining and interactive predictions needed to make an experience like this work is difficult for existing machine learning infrastructure. Learn how we approach these challenges by using a technique common within database technology called "incremental execution."

Chapters
0:00 - Intro
0:15 - Dealing with unstructured data in PDFs
1:06 - Use case: Upload unstructured docs directly to the database
2:00 - How to solve this with Impira
3:15 - Why is this necessary though?
4:13 - Event processing: Extraction
5:59 - File upload
6:26 - Incremental execution in Impira
7:04 - IQL (Impira Query Language)

Relevant links:

--