AI Center

DELIVERY:

Last updated Sep 1, 2025

Use Custom NER with continuous learning

Background information

This example is used to extract chemicals by the category mentioned in research papers. By following this procedure you will extract the chemicals and categorize them as ABBREVIATION, FAMILY, FORMULA, IDENTIFIER, MULTIPLE, SYSTEMATIC, TRIVIAL and NO_CLASS.

Prerequisites

This procedure uses the Custom Named Entity Recognition package. For more information on how this package works and what it can be used for, check the Custom Named Entity Recognition page.

For this procedure, we have provided sample files as follows:

Pre-labeled training dataset in CoNLL format. You can download the training dataset from the following link: training dataset.
Pre-labeled test dataset. You can download the test dataset from the following link: test dataset.
Sample workflow for extracting categories of chemicals mentioned in research papers. You can download it from the following link: sample workflow.
Note: Make sure that the following variables are filled in in the sample file:
- in_emailAdress - the email address to which the Action Center task will be assigned to
- in_MLSkillEndpoint - public endpoint of the ML Skill
- in_MLSkillAPIKey - API key of the ML Skill
- in_labelStudioEndpoint - optional, to enable continuous labeling: provide import URL of a label studio project

Procedure

Use the following steps to extract chemicals by their category from research papers.

Import the sample dataset in UiPath® AI Center. To do so, go to the Datasets menu and upload the train and test folder from the sample.
Select the desired Custom NER package from ML Packages > Out of the Box Packages > UiPath Language Analysis and create it.
Go to the Pipelines menu and create a new full pipeline run for the package created in the previous step. Point to the training and the test dataset provided in the sample file.

Once complete, the results of the pipeline are available under Artifacts. You can download the classification report to evaluate the test results.
Create a new ML Skill using the package generated by the pipeline run from the previous step and deploy it.
Once the skill is deployed, leverage it in the UiPath® Studio workflow provided. To enable capturing data with weak predictions, deploy a Label Studio instance and provide the instance URL and API key in the Label Studio activity in the workflow.

Getting started with Label Studio

Install Label Studio on your local machine or cloud instance. To do so, follow the instructions from the Label Studio page.
Create a new project from the Named Entity Recognition Template and define your Label Names.
Make sure that the label names have no special characters or spaces. For example, instead of Set Date, use SetDate.
Make sure that the value of the <Text> tag is "$text".
Upload the data using the API from the Label Studio API page.
cURL request example:
```
curl --location --request POST 'https://<label-studio-instance>/api/projects/<id>/import' \)\)
--header 'Content-Type: application/json' \)\)
--header 'Authorization: Token <Token>' \)\)
--data-raw '[
    {
      "data": {
        "text": "<Text1>"
      },
    },
    {
      "data": {
        "text": "<Text2>"
       }
    }
]'curl --location --request POST 'https://<label-studio-instance>/api/projects/<id>/import' \)\)
--header 'Content-Type: application/json' \)\)
--header 'Authorization: Token <Token>' \)\)
--data-raw '[
    {
      "data": {
        "text": "<Text1>"
      },
    },
    {
      "data": {
        "text": "<Text2>"
       }
    }
]'
```
Annotate your data.
Export the data in CoNLL 2003 format and upload it to AI Center.
Provided the Label Studio instance URL and API key in the provided sample workflow in order to capture incorrect and low confidence predictions.