New paper: Isolating creative vs. algorithm effects in ad A/B tests via three-arm design
asusarla · x · 2026-08-25
A new working paper proposes a three-arm experimental design to disentangle creative performance from platform algorithm bias in ad A/B testing. In addition to control and standard creative arms, it adds an arm where the treatment creative is visible only to the algorithm, not users, allowing researchers to isolate the algorithm's preference for specific creatives.
More from Research
- GPT-5.6 achieves ~10% success rate on 3,300 open math problems — littmath · 2026-08-25
- SSR: Self-Speculation for Faster Reasoning Models — kastnerkyle · 2026-08-25
- Training AI to Paint with Code via Reinforcement Learning — kastnerkyle · 2026-08-25
- WorldMind: Decoupled Game World Model for State-Aware NPC Behavior — Zhiyang Deng · 2026-08-25
- Nature Publishes ML-Based Probabilistic Weather Forecasting — pastramimachine · 2026-08-25
- Moderna reveals how it uses AI to accelerate mRNA vaccine development — alex_verem · 2026-08-25