GenPage: Towards End-to-End Generative Homepage Construction at Netflix
GenPage is a generative model developed by Netflix to create personalized homepages by leveraging user context and reinforcement learning, resulting in improved user engagement and reduced latency compared to traditional multi-stage recommendation systems.
MAIN POINTS
- GenPage generates Netflix homepages using a single transformer model, replacing complex multi-stage systems.
- The model uses user context as a prompt to autoregressively generate personalized homepages.
- Reinforcement learning optimizes page-level rewards, enhancing diversity and user satisfaction.
- Custom tokenization and hybrid row decoding improve efficiency and enforce business rules.
TAKEAWAYS
- GenPage's generative approach significantly improves user engagement and reduces serving latency by 20%.
- Enriching user context yields better personalization than merely scaling model capacity.
- Reinforcement learning increases homepage diversity, indicating effective page-level optimization.
- The model's flexibility allows for easy adaptation to new content types and personalization features.