Now available, in Oracle Cloud Infrastructure (OCI) Language version 4.1 is now available and introduces powerful enhancements to our AI language service, enabling organizations to use their data with exceptional accuracy, security, and performance.

Enhanced accuracy for securing sensitive data

In today’s data-driven world, protecting personally identifiable (PII) is crucial. The OCI Language PII service is now even more accurate at identifying sensitive information and protected health information (PHI). A top customer request has been the ability to customize PII results for specific business needs, and Version 4.1 introduces two features that make this capacity possible. Now, you can create custom detection rules based in regular expressions (REGEX) to fine-tune entity detection to unique requirements. Additionally, the new custom rules engine allows organizations to exclude specific entity types from PII or PHI detection, offering flexibility to target only relevant information and reduce false positives.

Figure 1: Custom PII detection rules
Figure 1: Custom PII detection rules
Figure 2: Exclude entity types within PII
Figure 2: Exclude entity types within PII

Translation services: Unparalleled control and performance

OCI Language translation services have undergone significant enhancements to provide organizations with greater control, performance, and isolation. With the ability to deploy on a dedicated endpoint, businesses can now enjoy improved translation accuracy, faster processing times, and enhanced security. This feature is particularly beneficial for organizations that require high levels of control and customization for their translation workflows.

We’ve also added some quality-of-life improvements for developers such as auto language detection in our document language API and enhanced support for popular file formats like SRT, JSON, and HTML. These features allow you to focus on business-critical projects rather than losing time formatting files.

We’ve also updated the translation models for 11 languages, including Arabic, Russian, Japanese, Korean, Simplified Chinese, Spanish, French, Hebrew, Portuguese, Brazilian Portuguese, German, and Traditional Chinese. These updated models provide organizations with the most accurate translations possible, enabling them to communicate effectively with global audiences.

API improvements for seamless integration

Our API improvements are designed to make integration seamless and efficient. With the introduction of alias names for endpoints, organizations can now update endpoints without requiring code changes as shown in Figure 3. This improvement helps reduce the administrative burden associated with endpoint updates, freeing up developers to focus on more strategic initiatives.

Create new endpoints without code
Figure 3: Creating a dedicated translation endpoint alias

Conclusion

OCI Language version 4.1 marks a major step forward in delivering a more accurate, secure, and high-performing language service available. With improved accuracy in detecting sensitive data, enhanced control and performance for translation, and streamlined API integration, this release is set to improve how organizations handle language processing. With these new features we continue to deliver high standards of accuracy and security empowering organizations to achieve greater efficiency, productivity, and innovation.

Additional exciting developments are in the works for our next version—stay tuned! And as always, we’d love to hear your feedback.

For more information, see the following resources: