Amazon Comprehend
Developer Guide

The AWS Documentation website is getting a new look!
Try it now and let us know what you think. Switch to the new look >>

You can return to the original look by selecting English in the language selector above.

Comprehend Custom

Customize Comprehend for your specific requirements without the skillset required to build machine learning-based NLP solutions. Using automatic machine learning, or AutoML, Comprehend Custom builds customized NLP models on your behalf, using data you already have. Training and calling custom comprehend models are both async (batch) operations.

Amazon Comprehend uses a proprietary, state-of-the-art sequence tagging deep neural network model that powers the same Amazon Comprehend detect entities service to train your custom entity recognizer models. In addition, we understand that acquiring training data could be costly. To help customers build a highly accurate model with limited amount of data, Amazon Comprehend uses a technique called transfer learning to train your custom models based on an sophisticated general-purpose entities recognition model that was pre-trained with a large amount of data we collected from multiple domains. Offline experiments showed that transfer learning significantly improved custom entity recognizer model accuracy especially when the amount of training data is small.