DoubleVerify Launches Semantic Science Division to Bolster Brand Safety & Performance Solutions

By DoubleVerify On Oct 9, 2019

What does “park” mean? Is it a verb or a noun? Is a “crane” construction equipment? Or is it a bird? Is “sex” a physical activity or male/female categorization? Usually, it is the context of how a word is used that conveys the true meaning.

In a world full of words, phrases and acronyms with multiple interpretations, it may be challenging for automated systems to determine the correct meaning of online copy and text. There have been significant advances in machine learning and artificial intelligence over the years. However, the semantic analysis of text — a linguistic approach for extracting the most appropriate contextual-usage meaning of words based on syntactic structures, phrases, clauses, sentences, paragraphs, and more — is a subject most ad tech and verification companies are just beginning to explore.

This is why, DoubleVerify is excited to officially announce the launch of our Semantic Science division. The Semantic Science division will combine AI textual analysis with deep human insight and expertise to understand the “meaning” of language so that we may classify it at an even more granular level.

Artificial Intelligence

Core to the Semantic Science division, DV is integrating the technology gained through the recent acquisition of contextual intelligence and content classification platform, Leiki. Leiki’s semantic AI software engine provides high-definition analysis of any piece of text — a complex news article, for example, or contextual video data. DoubleVerify then uses this analysis data to protect brand reputation throughout the media transaction (pre-and post-bid) and to enable proactive contextual targeting of content aligned with a brand’s equity or target audience profile. DV is the first verification company to leverage semantic AI to serve advertisers at scale.

Human Expertise

Based in Helsinki, DV’s Semantic Science division is comprised of dozens of language and software development specialists whose key competency is advanced content classification, based on semantic understanding of text. This team was responsible for developing one of the world’s most comprehensive general-purpose ontologies for Leiki — consisting of over 200,000 language-independent concepts and using more than 5 million rules for finding the correct meaning of each piece of content — which will now be integrated into the DV platform.

DV’s Semantic Science team provides the industry’s most accurate content classification system, ensuring advertisers appear beside appropriate and relevant media.

“In an era of rapid publishing cycles and increasingly divisive content, protecting brand reputation is critical to advertisers,” said Dan Slivjanovski, CMO, DoubleVerify.

“Just as our Fraud Lab detects emerging fraud schemes, our specialized Semantic Science team — combining AI and human expertise — will focus on understanding the context and true meaning of digital content. This supports our mission to power marketing performance, giving advertisers clarity and confidence in their digital investment.”

(This was republished with permission from DV)