End-to-End Inventory Prediction and Contract Allocation for Guaranteed Delivery Advertising

Our paper on advertising sales management will be presented at the KDD 2023. Thanks to all the authors: Wuyang Mao, Chuanren Liu, Yundu Huang, Zhonglin Zu, M Harshvardhan, Liang Wang, and Bo Zheng.


Guaranteed Delivery (GD) advertising plays an essential part in e-commerce marketing, where the ad publisher signs contracts with advertisers in advance by promising delivery of advertising impressions to fulfill targeting requirements for advertisers. Previous research on GD advertising mainly focused on online serving yet overlooked the importance of contract allocation at the GD selling stage. Traditional GD selling approaches consider impression inventory prediction and contract allocation as two separate stages. However, such a two-stage optimization often leads to inferior contract allocation performance. In this paper, our goal is to reduce this performance gap with a novel end-to-end approach. Specifically, we propose the Neural Lagrangian Selling (NLS) model to jointly predict the impression inventory and optimize the contract allocation of advertising impressions with a unified learning objective. To this end, we first develop a differentiable Lagrangian layer to backpropagate the allocation problem through the neural network and allow direct optimization of the allocation regret. Then, for effective optimization with various allocation targets and constraints, we design a graph convolutional neural network to extract predictive features from the bipartite allocation graph. Extensive experiments show that our approach can improve GD selling performance compared with existing two-stage approaches. Particularly, our optimization layer can outperform the baseline solvers in both computational efficiency and solution quality. To the best of our knowledge, this is the first study to apply the end-to-end prediction and optimization approach for industrial GD selling problems. Our work has implications for general prediction and allocation problems as well.

Decision Aggregation with Reliability Propagation

Congratulations to Yuyue Chen (Drexel University); Hao Zhong (ESCP Business School), Churanren Liu (University of Tennessee, Knoxville), Hande Benson (Drexel University) for their article on “Decision Aggregation with Reliability Propagation”, winning the Best Paper Runner-Up Award at WITS 2022: https://witsconf.org/wits2022-awards/


People often make decisions differently, even when faced with the same decision-making scenario and objectives, due to their varying abilities to access, process, and comprehend information relevant to the decisions at hand. To reconcile these differing perspectives and arrive at a unified decision, various approaches such as crowd-sourced systems have been developed to tap into the collective intelligence embodied in the opinions from a group of individuals. The diversity of opinions is both cure and curse for the effective use of crowd-sourced intelligence. To unify crowd-sourced intelligence for a well-informed decision, we propose an algorithmic approach for decision aggregation that accurately quantifies the reliability of information from multiple sources. The key idea behind this approach is to model the propagation of reliability in decisions based on an ensemble of relevance graphs, where the optimization of both the reliability propagation and the graph ensemble are mutually reinforced. The propagated reliability can be used to aggregate intelligence from multiple sources and facilitate decision-making by leveraging various types of inter-correlations of information sources and the subjects of the information. Meanwhile, the optimized graph ensemble can retain the relevance structures with respect to the crowd-sourced intelligence. We evaluate our approach with large-scale data sets, and the results show that, when aggregating analysts’ recommendations in stock markets, our approach not only outperforms alternative methods, but also provides interesting insights into the reliability of the information.