Use imported ML models to find custom terms

Rule-based model example

For example, when a machine learning model is applied as an enrichment to a field, it extracts all entity types in that field that were specified in a Knowledge Studio rule-based model. If the model recognizes entity types such as person, surname, and job title they are recognized in your documents and tagged.

In the output, the information that is extracted by the Machine Learning enrichment in the enriched_{field_name} array, within the entities array. In this example, the field that is selected for enrichment is text.

{
  "enriched_text": [
    {
      "entities": [
        {
          "path": ".wksrule.entities.PERSON",
          "text": "George Washington",
          "type": "PERSON"
        },
        {
          "path": ".wksrule.entities.GIVENNAME",
          "text": "George",
          "type": "GIVENNAME"
        },
        {
          "path": ".wksrule.entities.SURNAME",
          "text": "Washington",
          "type": "SURNAME"
        },
        {
          "path": ".wksrule.entities.POSITION",
          "text": "politician",
          "type": "POSITION"
        },
        {
          "path": ".wksrule.entities.POSITION",
          "text": "soldier",
          "type": "POSITION"
        },
        {
          "path": ".wksrule.entities.JOBTITLE",
          "text": "President of the United States",
          "type": "JOBTITLE"
        }
      ],
      "text": [
        "George Washington (February 22, 1732‚ December 14, 1799) was an American politician and soldier who served as the first President of the United States from 1789 to 1797 and was one of the Founding Fathers of the United States."
      ]
    }
  ]
}

As a result, if someone uses the API to submit a Discovery Query Language query to look for occurrences of the enriched_{field_name}.entities.type:jobtitle enrichment, any passages that discuss a person's job title are returned.

Machine learning model example

In this example, a Machine learning model extracts entity types such as person, oranization, and date, and information about relationships between the entities. When this ML model is applied as an enrichment to a field, it uses machine learning to understand the linguistic nuances, meaning, and relationships that are mentioned in the document.

In the output, the information that is extracted by the Machine Learning enrichment in the enriched_{field_name} array, within the entities and the relations arrays. In this example, the field that is selected for enrichment is text.

{
  "enriched_text": [
    {
      "entities": [
        {
          "count": 1,
          "text": "Democratic Party",
          "type": "ORGANIZATION"
        },
        {
          "count": 1,
          "text": "March 15, 1767",
          "type": "DATE"
        },
        {
          "count": 1,
          "text": "President",
          "type": "POSITION"
        },
        {
          "count": 1,
          "text": "Andrew Jackson",
          "type": "PERSON"
        }
      ],
      "relations": [
        {
          "sentence": "Andrew Jackson (March 15, 1767‚ June 8, 1845) was an American soldier and statesman who served as the seventh President of the United States from 1829 to 1837 and was the founder of the Democratic Party."
        }
      ]
    }
  ]
}

Machine learning model limits

The number of Machine Learning (ML) models you can create per service instance depends on your Discovery plan type.

For each Knowledge Studio machine learning model, the maximum number of entities that can be detected is 50.

Advanced rules models

Add an advanced rules model to apply a text extraction model that was created and exported from the Advanced Rule editor of IBM Watson® Knowledge Studio to your collection.

Your model must be created with the appropriate Knowledge Studio deployment:

• IBM Cloud Pak for Data You can add models that were created and exported from the following places:

  • IBM Watson® Knowledge Studio that was built with a IBM Cloud Pak® for Data deployment earlier than the 4.5 release.
  • IBM Watson® Knowledge Studio that is hosted on IBM Cloud
  • NLP Editor that is built by contributors to the Center for Open-source Data & AI Technologies

• IBM Cloud You can add models that were created with a IBM Watson® Knowledge Studio instance that is hosted on IBM Cloud only.