Introduction

This live lab demonstrates how to extract insights from unstructured text with no machine-learning knowledge. OCI Language Service provides suite of services to distill a deeper understanding of opinions with sentiment analysis, identify key phrases and extract named entities such as people, places and organizations to understand common subjects and patterns. You can use out of the box pre-trained models and also customize the models to suite a specific domain.

OCI Language Service key features:

Pre-Trained Models

OCI Language pre-trained APIs uses AI models trained for most common use cases.

1. Sentiment Analysis

Identifies sentiment at the document, sentence and aspect level.

2. Named Entity Recognition

Identifies common entities, people, places, locations, email, and so on.

3. Key Phrase Extraction

Identify the most salient talking points in your text.

4. Language Detection

Detects languages based on the given text, and includes a confidence score.

5. Text Classification

Identifies the document category and subcategory that the text belongs to.

6. Personal Identifiable Information

Detects various entities of personal information.

7. Personal Health Information

Detects various entities of personal health information.

Custom Models

OCI Language custom models enables you to customize Text Classification and Named Entity Recognition with your own data

1. Custom Text Classification

Enables you to build a custom AI model to automatically classify text into a set of classes you pre-define, both single label and multi-label custom text classification is supported.

2. Custom Named Entity Recognition

Enables you to build a model to identify domain-specific entities that are unique to your business or industry vertical.

Text Translation

OCI Language now provides an API to automatically translate text across 21 languages.

Jobs

OCI Language now provides an API to create async jobs that enables you to process large volumes of text and translate documents in a batch mode.

Healthcare models

Provides state of the art service to extract entities from healthcare records.

Multilingual models

Provides support for using machine learning models with non-English input content.

Estimated Lab Time: 2 hour

Objectives:

• Understand overview of the OCI Language
• Learn how to analyze text using pre-trained models using OCI Console, API and SDK
• Learn how to create and train custom models
• Learn how to use OCI Language service to translate text using OCI Console, API and SDK
• Learn how to use OCI Language service to create jobs using OCI Console, API and SDK

Prerequisites:

• An Oracle Free Tier, or Paid Cloud Account
• Additional prerequisites (cloud services) are mentioned per lab
• Familiar with OCI Policy and SDK/CLI setup.
• Familiar with Python/Java programming for SDK usage is strongly recommended.
• Familiar with OCI services like DataScience, DataFlow etc., are recommended, but not required.
• Familiar with editing tools (vim, nano) or shell environments (cmd, bash, etc) (Optional for API integration)

Proceed to the next section

×