PYLER: Solving Challenges in the Digital Advertising Market with AI

PYLER's AiD safeguards brands in the risky world of video advertising with AI-powered contextual targeting.
Ranked 2nd globally at CVPR 2022, PYLER proves its cutting-edge capabilities in video understanding AI.

Launching AiD – Protecting Brand Safety in the Era of Video Advertising

PYLER addresses the growing brand safety concerns in the digital video advertising space with world-class Video Understanding AI, offering contextual video marketing solutions tailored for a new era.

In 2022, PYLER proved its technological prowess at CVPR, the world’s top AI and computer vision conference, by achieving outstanding results alongside global tech giants such as Intel, Tencent, and ByteDance. Through proprietary computer vision technologies and a brand safety algorithm built on IAB (Interactive Advertising Bureau) standards, Pyler has emerged as a highly competitive player.

Recognizing the seriousness of brand safety issues in video advertising, PYLER has launched AiD, an AI-powered solution to tackle these challenges head-on. As a leader in brand safety and contextual targeting in Korea, PYLER aims to raise industry awareness and restore an advertising environment where great ads are matched with trustworthy content.

Chapter 1:
Ranked 2nd Globally at CVPR 2022

PYLER was invited as a finalist to CVPR 2022, the world's largest conference on computer vision and pattern recognition, and proudly secured 2nd place in Track 3 of the LOVEU (Long-form Video Understanding) Workshop.

This challenge tested how effectively AI models could learn from instructional videos and scripts, and guide users through step-by-step tasks. In simple terms, the AI was required to interpret tutorial videos and then provide helpful answers when a user asked questions — similar to testing how well the AI "understands" the manual.

PYLER’s multimodal model—which integrates video, image, and text data—was praised for its ability to align complex questions with context and deliver user-relevant responses in sequential steps, making the task particularly challenging.

Dongchan Park (CTO) and Sangwook Park (ML Lead) of PYLER’s AI Context Lab shared:

“We competed against leading global corporations and research institutions with significantly fewer resources. Achieving 2nd place was meaningful—especially since we ranked 1st in Recall@3, one of the key evaluation metrics.”

Chapter 2:
AiD – Rebuilding Trust in Video Advertising

Brand safety concerns in platforms like YouTube have become severe across global markets, including South Korea. According to Pixability, one-third of total ad spend in 2021 was unintentionally spent on harmful or inappropriate content.

In one notable incident, KOBACO (Korea's only public ad agency) placed public service announcements on adult-themed YouTube channels, sparking controversy. Even major Korean corporations are often unaware of the exact content their ads are appearing alongside.

PYLER identified this market inefficiency—where 16% to 37% of ad spend is wasted on brand-damaging content—and created AiD: a 24/7 AI-powered guardian that continuously analyzes video content, flags risky categories, and blocks harmful ad placements.

Source: 2021 Pixability

While global advertisers now prioritize brand safety even above performance, many Korean advertisers remain uncertain about how to address the issue. PYLER hopes AiD will help normalize the ad ecosystem by giving advertisers peace of mind.

In the long run, AiD also aligns with privacy-first trends. It leverages contextual targeting without relying on third-party cookies, delivering efficient and effective ad performance without compromising user privacy.