Amazon Comprehend

Enable your integration with models from Amazon Comprehend.

We have simplified the process of connecting our system with the Amazon Comprehend model. If you already have data and training pipelines in Amazon Comprehend, you can utilize your trained model to enhance the performance.

To use the Amazon Comprehend model, follow these steps:

Setup Datasaur ML Assisted Extension

Please make sure you have access to our ML Assisted Extension with AWS Comprehend as a provider.

  1. Create a custom project for Row Labeling

  2. Click "Manage Extension" on your right bar.

  3. Pop Up Manage Extension will appear and you can enable the Datasaur ML Assisted.

  1. Datasaur ML Assisted is now enabled and select Amazon Comprehend as Provider

Setup Role in AWS Identity and Access Management (IAM)

Before we start to create Role in the IAM Page, please create these policies first for your permissions when creating a new Role.

  1. Go to your IAM Page, navigate to the Policies section, and then create a new policy.

  2. You can download the examples below.

Now you can continue to create a new Role with the policy.

  1. Go to the IAM Page and navigate to the Roles section

  2. Click on "Create Roles" then select "AWS Account" as the role type and insert the Account ID and External ID. These values are automatically generated within the Datasaur ML Assisted and can be copied and pasted into the AWS platform.

  1. After finishing the first step, you can see the "Add permissions" step and search the policy you created before. Add them to the permissions.

  1. Provide a role name and click “Create role.”

  2. The role will be created successfully.

  3. View your role and copy the Role ARN.

  4. Paste this information into the ML-Assisted Labeling extension.

  5. Navigate to the Amazon Comprehend page and retrieve the Endpoint ARN. Copy the obtained Endpoint ARN and paste it into the extension.

Start Prediction

Once you have configured the above options, you can predict labels and obtain predicted labels from your own model by clicking “Predict Labels”.

Last updated