{"id":861,"date":"2021-09-17T06:56:44","date_gmt":"2021-09-17T06:56:44","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2021\/09\/17\/custom-document-annotation-for-extracting-named-entities-in-documents-using-amazon-comprehend\/"},"modified":"2021-09-17T06:56:44","modified_gmt":"2021-09-17T06:56:44","slug":"custom-document-annotation-for-extracting-named-entities-in-documents-using-amazon-comprehend","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2021\/09\/17\/custom-document-annotation-for-extracting-named-entities-in-documents-using-amazon-comprehend\/","title":{"rendered":"Custom document annotation for extracting named entities in documents using Amazon Comprehend"},"content":{"rendered":"
\n

Intelligent document processing (IDP), as defined by IDC, is an approach by which unstructured content and structured data is analyzed and extracted for use in downstream applications. IDP involves document reading, categorization, and data extraction, by using AI\u2019s processes of computer vision (CV), Optical Character Recognition (OCR), and natural language processing (NLP) on provided texts.[1] Companies in financial services, healthcare, and manufacturing process millions of documents each year and find the process to be painstaking. Document processing is manual, slow, expensive, and error-prone, and data is often spread across disparate sources. As a result, creating and managing a document processing pipeline remains a challenge for many companies.<\/p>\n

Amazon Comprehend<\/a> is an NLP service that provides APIs to extract key phrases, contextual entities, events, and sentiment from documents. Entities<\/em> refer to things in your document such as people, places, organizations, credit card numbers, and so on. But what if you want to identify entity types unique to your business, like proprietary product codes or industry-specific terms? Custom entity recognition in Amazon Comprehend enables you to train models to extract entities that are unique to your business in just a few easy steps. Historically, you could only use Amazon Comprehend APIs on plain text documents. If you wanted to process Word or PDF documents, you first needed to preprocess or flatten the documents into a plain text format, which often reduces the quality of the context within the document.<\/p>\n

Today, Amazon Comprehend launched a feature to help organizations extract insights and automate processing documents of different formats (PDF, Word, plain text) and layouts (bullets, lists). You can now use custom entity recognition in Amazon Comprehend directly on more document types (scanned PDF, machine-readable PDF, and Word documents) without needing to convert your files to plain text. Custom entity recognition can now extract custom entities and process varying document layouts such as dense text and lists or bullets in PDF, Word, and plain text documents\u2014at no additional cost.<\/p>\n

To train a custom entity recognition model that you can use on your PDF, Word, and plain text documents, you need to first annotate PDF documents using a custom\u00a0Amazon SageMaker Ground Truth<\/a> annotation template provided by Amazon Comprehend. The custom template renders a PDF and allows you to annotate directly on the document. The custom model leverages both the natural language and structural or positional information (coordinates) of the text to make sure that we accurately extract custom entities that might have been previously impacted when flattening a document.<\/p>\n

In this post, we walk through the steps of setting up the custom annotation template and show examples of how to annotate finance documents (SEC filings). After you annotate your data, you can use the generated manifest file to train a custom entity recognition model.<\/p>\n

Feature overview<\/h2>\n

We cover the following steps:<\/p>\n

    \n
  1. Install the annotation tool.<\/li>\n
  2. Upload training documents.<\/li>\n
  3. Create a private workforce to label the training documents.<\/li>\n
  4. Set up the Ground Truth labeling job.<\/li>\n
  5. Access the annotation login portal to start labeling documents.<\/li>\n
  6. Label documents with the custom template.<\/li>\n<\/ol>\n

    Only PDF files are used for training a custom entity recognition model, but you can use that model on plain text, Word, and PDFs documents during inference (for now, asynchronous processing only).<\/p>\n

    The custom annotation template is available to download on GitHub<\/a>. You need to Git clone or download this application and install it in local or in an AWS Cloud 9<\/a> integrated development environment (IDE). AWS Cloud9 is a cloud-based IDE that lets you write, run, and debug your code with just a browser. This annotation tool needs to be built and deployed using Python and AWS Serverless Application Model<\/a> (AWS SAM) (the AWS SAM CLI lets you locally build, test, and debug serverless applications that are defined by AWS SAM templates).<\/p>\n

    Prerequisites<\/h2>\n

    You need to install the following to build and deploy this solution:<\/p>\n

      \n
    1. Install<\/a> Python 3.8.x.<\/li>\n
    2. Install<\/a> jq.<\/li>\n
    3. Install<\/a> the AWS SAM CLI.<\/li>\n
    4. Make sure you have pip installed<\/a>.<\/li>\n
    5. Install and configure<\/a> the AWS CLI.<\/li>\n
    6. Configure<\/a> your AWS credentials.<\/li>\n<\/ol>\n

      If you\u2019re setting up this tool in us-east-1<\/code>, you can skip installing the AWS CLI and AWS SAM CLI, because it\u2019s already installed with Python environment. You just need to create a virtual environment<\/a> to use Python 3.8 in AWS Cloud9.<\/p>\n

      Install the annotation tool<\/h2>\n

      Download the annotation tool from GitHub<\/a>. You can either Git clone this repository or download the setup as a .zip file in your local. The .zip file contains all the installation parts required to build out the custom annotation workflow to annotate PDFs in Ground Truth.<\/p>\n

      First, you run and configure this tool from the AWS CLI. With AWS SAM, we build and deploy this tool to create an AWS CloudFormation<\/a> template. The CloudFormation template configures resources such as an AWS Lambda<\/a> function, AWS Identity and Access Management<\/a> (IAM) roles and permissions to create an Amazon SageMaker<\/a> labeling job, and your Amazon Simple Storage Service<\/a> (Amazon S3) bucket.<\/p>\n

      Once you unzip the tool, you need to run commands to build and deploy from with-in the annotation tool. In the CLI, navigate to folder ComprehendSSIEAnnotationTool<\/code> or run<\/p>\n

      cd ComprehendSSIEAnnotationTool<\/code> to change your directory to the annotation tool folder.<\/p>\n

        \n
      1. Run the following steps to build the tool:<\/li>\n<\/ol>\n

        make bootstrap<\/code><\/p>\n

        make build<\/code><\/p>\n

          \n
        1. After the build is successful, run the following command to deploy:<\/li>\n<\/ol>\n

          make deploy-guided<\/code><\/p>\n

            \n
          1. For Stack Name<\/strong>\u00b8 enter comprehend-annotation-tool<\/code>.<\/li>\n
          2. For AWS Region<\/strong>, enter the Region you\u2019re in.<\/li>\n
          3. Leave the remaining parameters as default.<\/li>\n
          4. For Confirm changes before deploy<\/strong>, enter Y<\/code>.<\/li>\n<\/ol>\n

            \"\"<\/p>\n

              \n
            1. For Allow<\/strong> SAM CLI IAM role creation<\/strong>, enter Y<\/code>.<\/li>\n<\/ol>\n

              After you complete these steps, the AWS CloudFormation resources are deployed in your AWS account, which you use to set up the annotation jobs. If deployment fails, delete the CloudFormation stack via the AWS CloudFormation console and try make deploy-guided<\/code> again from your local or AWS Cloud9 IDE to reinstall this tool.<\/p>\n

                \n
              1. When the status of the CloudFormation deployment changes from In-Progress<\/code> to Complete<\/code> (which usually takes a few minutes), note the S3 bucket name created by the CloudFormation template.<\/li>\n<\/ol>\n

                You need this S3 bucket to upload training documents that are used for annotation in the next step.<\/p>\n

                \"\"\u00a0<\/strong><\/p>\n

                Upload the training documents<\/h2>\n

                You can run the following command in your local to copy local files to the S3 bucket created in the previous step. As a reminder, these training documents should be PDF documents only.<\/p>\n

                \n
                aws s3 cp --recursive <local-path-to-source-docs><\/span><\/em> s3:\/\/comprehend-semi-structured-documents-${AWS_REGION}-${AWS_ACCOUNT_ID}\/source-semi-structured-documents\/<\/code><\/pre>\n<\/p><\/div>\n
                Or you can directly upload documents to this S3 bucket created by the CloudFormation template by creating a folder source-semi-structured-documents<\/code>. Refer to this link<\/a> to learn how to upload files in Amazon S3 and creating folders.<\/p>\n
                Create a private workforce<\/h2>\n
                You need to create a private workforce in Ground Truth, which labels the training documents you uploaded in Amazon S3. You can view or create a private workforce using the Ground Truth console. For detailed instructions, see the section \u201cCreating a private work team\u201d in Developing NER models with Amazon SageMaker Ground Truth and Amazon Comprehend<\/a>. To follow along with the next steps, we recommend adding yourself to the private workforce. All new workers added to your private workforce receive an enrollment email with the URL, user name, and a temporary password to log in to the labeling portal. Your workers use this login information after we create the labeling job.<\/p>\n
                After you create a private workforce, note the private workforce name, which we need when creating a Ground Truth labeling job using the AWS CLI.<\/p>\n
                Set up the Ground Truth labeling job<\/h2>\n
                Ground Truth allows you to create custom labeling workflows, which you can use when the provided templates don\u2019t suit the requirements for your labeling efforts.<\/p>\n
                To create a Ground Truth labeling job, you need to set up parameters such as the input Amazon S3 path, CloudFormation stack name, work team name, Region, job name prefix, entity types, and annotator metadata.<\/p>\n