TSPA Member Presentation by ActiveFence – Data Centric AI: Robust and Accurate AI for Content Moderation

Hosted by TSPA

Data centric AI—an approach that prioritizes quality and diversity of data over complex model architecture—is key for enhancing the effectiveness of language models for classification. In this talk, Matar Haller, VP Data & AI at ActiveFence, will discuss how ActiveFence uses a data centric approach in the development of robust, accurate, and unbiased language models for content moderation, illustrated by real-world examples. Matar will also explain how these models are critical in detecting violations in real time at scale, across a variety of abuse areas, including hate speech, harassment, and adult content, thus ensuring the safety of online interactions. This TSPA member expert presentation is open to everyone, but registration is required.