Sinequa enhances NLP and data connectors

Sinequa, a provider of enterprise search solutions, has added new capabilities in its platform. With an expansion of its natural language processing (NLP) capabilities, Sinequa has introduced additional support for named entity recognition while new connectors expand the range of systems and formats that can be ingested.

Additionally, there is support for named entity recognition packaged within the platform, such as weights and measures and personally identifiable information (PII) across different geographies. There are also new capabilities to extend the platform’s ability to automatically conduct intelligent classification across vast amounts of indexed enterprise content and data.

“With the latest release, Sinequa can better serve the needs of customers in various industries like life sciences, manufacturing, and finance, which requires recognizing concepts and highly technical and specific vocabularies,” said Philippe Motet, vice president of engineering at Sinequa. “We are improving the efficiency and speed at which employees can find relevant information and insights needed to make smarter and more effective business decisions.”

With expanded NLP capabilities, this new release recognizes PII, which enables instant identification of 24 categories of PII,  allowing organizations to comply with privacy-protection legislation such as GDPR more effectively. These entities cover 89 types within these categories, including BBAN, IBAN, credit card number, date of birth, driver’s license number for 13 countries, license plates for 11 countries, passport number for 13 countries, and various national identification numbers for 14 countries.

Building on Sinequa’s support for production-scale machine learning, the latest release provides several new, interrelated capabilities to help organizations attain the benefits of AI-powered intelligent search more quickly and without the need for data science skills or machine learning expertise. According to Sinequa, all tasks concerned with model building and validating and keeping accurate predictions over time by adjusting to concept and data drift are now supported within the platform.

Sinequa's Intelligent Labeling Application is dedicated to the data management of the classification algorithm. A critical design goal of this new application is to provide a user experience that compels subject matter experts to use the application and thereby provide their domain expertise. This enables organizations to create an objective training set without relying on data scientists, while also allowing them to leverage the feedback of subject matter experts to improve model predictions over time, while handling concept drift and data drift by automatically requesting their validation for ambiguous predictions.  

Adding to its growing family of 200-plus smart connectors, the latest version of the Sinequa platform features connectors to the IBM Filenet P8 V5 application as well as for the Perforce Helix core application used for software version control.

For more information, go to www.sinequa.com.