AliExpress Learning-To-Rank: Maximizing Online Model Performance without Going Online

Huzhang, Guangda; Pang, Zhen-Jia; Gao, Yongqing; Liu, Yawen; Shen, Weijie; Zhou, Wen-Ji; Da, Qing; Zeng, Anxiang; Yu, Han; Yu, Yang; Zhou, Zhi‐Hua

doi:10.48550/arxiv.2003.11941

Cited by 2 publications

(1 citation statement)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, the offline training objective (e.g., loss functions, ranking metrics) has a significant gap to the true business objective (e.g., number of purchases). Many practitioners report that a fine-tuned model with a better offline performance does not correspondingly improve the online performance [6][7][8][9][10]. Therefore, it is ideal to find an ensemble learning algorithm that can output diversified items and can directly optimize the online performance.…”

Section: Introductionmentioning

confidence: 99%

Learning-To-Ensemble by Contextual Rank Aggregation in E-Commerce

Wang¹,

Huzhang²,

Lin³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Ensemble models in E-commerce combine predictions from multiple sub-models for ranking and revenue improvement. Industrial ensemble models are typically deep neural networks, following the supervised learning paradigm to infer conversion rate given inputs from sub-models. However, this process has the following two problems. Firstly, the point-wise scoring approach disregards the relationships between items and leads to homogeneous displayed results, while diversified display benefits user experience and revenue. Secondly, the learning paradigm focuses on the ranking metrics and does not directly optimize the revenue. In our work, we propose a new Learning-To-Ensemble (LTE) framework RA-EGO, which replaces the ensemble model with a contextual Rank Aggregator (RA) and explores the best weights of sub-models by the Evaluator-Generator Optimization (EGO). To achieve the best online performance, we propose a new rank aggregation algorithm TournamentGreedy as a refinement of classic rank aggregators, which also produces the best average weighted Kendall Tau Distance (KTD) amongst all the considered algorithms with quadratic time complexity. Under the assumption that the best output list should be Pareto Optimal on the KTD metric for sub-models, we show that our RA algorithm has higher efficiency and coverage in exploring the optimal weights. Combined with the idea of Bayesian Optimization and gradient descent, we solve the online contextual Black-Box Optimization task that finds the optimal weights for sub-models given a chosen RA model. RA-EGO has been deployed in our online system and has improved the revenue significantly.

show abstract