I build the extraction pipelines that have to hold up on the whole corpus, not the slice in the demo. Which usually means I get called in after the celebration. The vendor demo ran clean: a dozen hand-picked PDFs went in, structured fields came out, the fields were correct, and everyone in the room exhaled. Somebody said the word “magic.” A six-figure number got attached to a roadmap. Then the pipeline met the real corpus — the ten thousand documents the business actually owns, the ones nobody curated — and the output quietly fell apart. Not with an error. With confident, plausible, wrong results that nobody caught until they were already downstream.
Transfer Learning is a powerful tool allowing us to use huge pre-trained models and tune them for our task. In this post we review a case of transferring weights of a CNN architecture trained ImageNet to detect chest X-Ray infected with pneumonia.