Zero-Shot Image Moderation in Google Ads with LLM-Assisted Textual Descriptions and Cross-modal Co-embeddings
This addresses the challenge of scalable and agile ads content moderation for Google, though it appears incremental as it builds on existing cross-modal and LLM techniques.
The paper tackles the problem of moderating massive volumes of diverse ads images at Google by proposing a zero-shot classification method using LLM-assisted textual descriptions and cross-modal co-embeddings, which significantly boosts detection of policy-violating content.
We present a scalable and agile approach for ads image content moderation at Google, addressing the challenges of moderating massive volumes of ads with diverse content and evolving policies. The proposed method utilizes human-curated textual descriptions and cross-modal text-image co-embeddings to enable zero-shot classification of policy violating ads images, bypassing the need for extensive supervised training data and human labeling. By leveraging large language models (LLMs) and user expertise, the system generates and refines a comprehensive set of textual descriptions representing policy guidelines. During inference, co-embedding similarity between incoming images and the textual descriptions serves as a reliable signal for policy violation detection, enabling efficient and adaptable ads content moderation. Evaluation results demonstrate the efficacy of this framework in significantly boosting the detection of policy violating content.