Causal-Informed Hybrid Online Adaptive Optimization for Ad Load Personalization in Large-Scale Social Networks
Aakash Mishra, Qi Xu, Zhigang Hua, Keyu Nie, Vishwanath Sangale, Vishal Vaingankar, Jizhe Zhang, Ren Mao
TL;DR
Problem: optimize ad load in a billion-user social network under business and user-experience constraints. Approach: a hybrid Primal-Dual + Bayesian Optimization framework (CTRCBO) that uses trust regions, local Gaussian Process surrogates for both objectives and constraints, and an upstream causal ML model to inform the surrogates, enabling efficient exploration-exploitation. Contributions: (i) a scalable CTRCBO algorithm, (ii) theoretical bounds showing per-trust-region regret $O(sqrt(T))$ and time-averaged constraint satisfaction ≤ 0, and (iii) validated improvements through online AB tests and real-world deployment on a billion-user platform. Significance: enables faster convergence, robust constraint adherence, and improved personalization metrics in large-scale online optimization settings.
Abstract
Personalizing ad load in large-scale social networks requires balancing user experience and conversions under operational constraints. Traditional primal-dual methods enforce constraints reliably but adapt slowly in dynamic environments, while Bayesian Optimization (BO) enables exploration but suffers from slow convergence. We propose a hybrid online adaptive optimization framework CTRCBO ( Cohort-Based Trust Region Contextual Bayesian Optimization), combining primal-dual with BO, enhanced by trust-region updates and Gaussian Process Regression (GPR) surrogates for both objectives and constraints. Our approach leverages a upstream Causal ML model to inform the surrogate, improving decision quality and enabling efficient exploration-exploitation and online tuning. We evaluate our method on a billion-user social network, demonstrating faster convergence, robust constraint satisfaction, and improved personalization metrics, including real-world online AB test results.