Adaptive Mixture Importance Sampling for Automated Ads Auction Tuning

TL;DR

This paper presents Adaptive Mixture Importance Sampling (AMIS), a novel method for optimizing KPIs in large-scale recommender systems and ad auctions, improving search diversity and convergence in noisy, dynamic environments.

Contribution

AMIS introduces a dynamic mixture proposal distribution that adjusts both parameters and mixing rates, outperforming traditional importance sampling methods in complex, noisy settings.

Findings

AMIS outperforms Gaussian Importance Sampling in offline simulations.

AMIS achieves better convergence and stability in noisy environments.

Online experiments show AMIS identifies more effective tuning points.

Abstract

This paper introduces Adaptive Mixture Importance Sampling (AMIS) as a novel approach for optimizing key performance indicators (KPIs) in large-scale recommender systems, such as online ad auctions. Traditional importance sampling (IS) methods face challenges in dynamic environments, particularly in navigating through complexities of multi-modal landscapes and avoiding entrapment in local optima for the optimization task. Instead of updating importance weights and mixing samples across iterations, as in canonical adaptive IS and multiple IS, our AMIS framework leverages a mixture distribution as the proposal distribution and dynamically adjusts both the mixture parameters and their mixing rates at each iteration, thereby enhancing search diversity and efficiency. Through extensive offline simulations, we demonstrate that AMIS significantly outperforms simple Gaussian Importance Sampling (GIS), particularly in noisy environments. Moreover, our approach is validated in real-world scenarios through online A/B experiments on a major search engine, where AMIS consistently identifies optimal tuning points that are more likely to be adopted as mainstream configurations. These findings indicate that AMIS enhances convergence in noisy environments, leading to more accurate and reliable decision-making in the context of importance sampling off-policy estimators.