Bid-Aware Active Learning in Real-Time Bidding for Display Advertising

Shuhao Liu,Yong Yu

doi:10.1109/access.2019.2961155

Abstract

In Real-time Bidding (RTB) based display advertising, demand side platforms (DSPs) estimate the click-through rate (CTR) of each advertisement impression, and then decide whether and how much to bid based on the information of the user and the advertiser. Typically, when a new campaign is launched, the CTR estimation module of the DSP needs to collect data to train an accurate estimator. The advertiser is charged for each ad impression in display advertising, therefore there is some cost for obtaining each training instance. Thus one crucial task is to actively train an accurate CTR estimator within the constraint of the budget. Traditional active learning algorithms fail to deal with such scenario because (i) acquiring training instances is implemented via performing real-time bidding for the corresponding auctions; (ii) RTB requires the bidding agent to make real-time decisions for sequentially coming bid requests; (iii) cost for each ad impression will be unveiled only after giving the bid price and winning the auction; (iv) training data gathered in post-bid stage has a strong bias towards the won impressions. In this paper, we propose a Bid-aware Active Real-time Bidding (BARB) algorithm to actively choose training instances by setting different bid prices for each ad auction, in order to efficiently train an accurate CTR estimation model within the budget constraint. The empirical study on different campaigns of three real-world datasets with three budget constraints shows the effectiveness of our proposed algorithm.

Highlights

Nowadays, Real-time Bidding (RTB) has become an important paradigm in display advertising [1]
Traditional active learning algorithms fail to deal with such scenario because (i) acquiring training instances is implemented via performing real-time bidding for the corresponding auctions; (ii) RTB requires the bidding agent to make real-time decisions for sequentially coming bid requests; (iii) cost for each ad impression will be unveiled only after responding a bid price and winning the auction; (iv) training data acquired in post-bid stage has a strong bias towards the won impressions
Based on the uniqueness of our problem, we propose an active learning framework for user response prediction in RTB, namely Bid-aware Active Real-time Bidding (BARB) to actively decide bid price for each ad impression

Summary

Introduction

Real-time Bidding (RTB) has become an important paradigm in display advertising [1]. Different from the conventional negotiation or pre-setting a fixed bid price for each campaign or keyword, RTB enables advertisers to give a bid price for every individual impression [2]. Each ad placement will trigger an auction when the user visits an ad-supported site (e.g., web page, streaming videos and mobile apps). The ad exchange will send bid requests [1] to the advertisers’ buying systems, usually referred to as Demand Side Platforms (DSPs). Upon receiving a bid request, a DSP will calculate a bid price as a response after holding an internal auction among all of its qualifying campaigns. An auction will be held at each intermediary

Methods

Results

Conclusion