Scalable hands-free transfer learning for online advertising

D'Alessandro, B; Chen, Daizhuo; Raeder, Troy; Perlich, Claudia; Williams, Melinda Han; Provost, Foster

doi:10.1145/2623330.2623349

Cited by 33 publications

(20 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Su et al [14] improves the click prediction by transferring information from the data-rich product to a data-scarce target product. Dalessandro et al [15] uses the web browsing data of the users as the source domain to predict user's engagement. They use the prior learned from the source domain as a regularizer for the logistic regression on the target domain.…”

Section: A Domain Adaptation and Transfer Learningmentioning

confidence: 99%

“…Aggarwal et al uses the re-targeting platform as the source domain and uses the large amount of re-targeting data to cold-start the partners in the prospecting platform, which is their target domain. Although our work addresses the same challenge as [15,16], we go a step further and use the data from the frequent-advertised partners to improve the performance on the target domain.…”

Section: A Domain Adaptation and Transfer Learningmentioning

confidence: 99%

See 1 more Smart Citation

Targeted display advertising: the case of preferential attachment

Manchanda

Yadav²,

Doan

et al. 2019

2019 IEEE International Conference on Big Data (Big Data)

View full text Add to dashboard Cite

An average adult is exposed to hundreds of digital advertisements daily 1 , making the digital advertisement industry a classic example of a big-data-driven platform. As such, the ad-tech industry relies on historical engagement logs (clicks or purchases) to identify potentially interested users for the advertisement campaign of a partner (a seller who wants to target users for its products). The number of advertisements that are shown for a partner, and hence the historical campaign data available for a partner depends upon the budget constraints of the partner. Thus, enough data can be collected for the high-budget partners to make accurate predictions, while this is not the case with the low-budget partners. This skewed distribution of the data leads to preferential attachment of the targeted display advertising platforms towards the high-budget partners. In this paper, we develop domain-adaptation approaches to address the challenge of predicting interested users for the partners with insufficient data, i.e., the tail partners. Specifically, we develop simple yet effective approaches that leverage the similarity among the partners to transfer information from the partners with sufficient data to cold-start partners, i.e., partners without any campaign data. Our approaches readily adapt to the new campaign data by incremental fine-tuning, and hence work at varying points of a campaign, and not just the coldstart. We present an experimental analysis on the historical logs of a major display advertising platform 2 . Specifically, we evaluate our approaches across 149 partners, at varying points of their campaigns. Experimental results show that the proposed approaches outperform the other domain-adaptation approaches at different time points of the campaigns.

show abstract

Section: A Domain Adaptation and Transfer Learningmentioning

confidence: 99%

Section: A Domain Adaptation and Transfer Learningmentioning

confidence: 99%

Targeted display advertising: the case of preferential attachment

Manchanda

Yadav²,

Doan

et al. 2019

2019 IEEE International Conference on Big Data (Big Data)

View full text Add to dashboard Cite

show abstract

“…By contrast, transfer learning methods may allow for arbitrary source and target tasks. In real time bidding based advertising, [Dalessandro et al, 2014] proposed a transfer learning scheme based on logistic regression prediction models, where the parameters of ad click prediction model were restricted with a regularisation term from the ones of user Web browsing prediction model. [Zhang et al, 2016a] extended it by considering matrix factorisation to fully explore the benefit of collaborative filtering.…”

Section: Transfer Learning From Web Browsing To Ad Clicksmentioning

confidence: 99%

Display Advertising with Real-Time Bidding (RTB) and Behavioural Targeting

Wang

Zhang

Yuan³

2017

FNT in Information Retrieval

View full text Add to dashboard Cite

The most significant progress in recent years in online display advertising is what is known as the Real-Time Bidding (RTB) mechanism to buy and sell ads. RTB essentially facilitates buying an individual ad impression in real time while it is still being generated from a user's visit. RTB not only scales up the buying process by aggregating a large amount of available inventories across publishers but, most importantly, enables direct targeting of individual users. As such, RTB has fundamentally changed the landscape of digital marketing. Scientifically, the demand for automation, integration and optimisation in RTB also brings new research opportunities in information retrieval, data mining, machine learning and other related fields. In this monograph, an overview is given of the fundamental infrastructure, algorithms, and technical solutions of this new frontier of computational advertising. The covered topics include user response prediction, bid landscape forecasting, bidding algorithms, revenue optimisation, statistical arbitrage, dynamic pricing, and ad fraud detection.

show abstract

“…For more details on our modeling methodology, see [14,4]. For all models, final decisions on segment selection and regularization strength were made by evaluating on a hold-out set with crisp labels created by probabilistic labeling.…”

Section: High-reach Online Demographic Targetingmentioning

confidence: 99%

“…Today we use very sophisticated optimization to make bidding decisions in 30ms that consider the consumer's web browsing history, search history, purchase patterns, social connections, and the context of the site where the ad is shown [13,14,4].…”

Section: Introductionmentioning

confidence: 99%

Pleasing the advertising oracle

Williams¹,

Perlich²,

D'Alessandro³

et al. 2014

Proceedings of the Eighth International Workshop on Data Mining for Online Advertising

Self Cite

View full text Add to dashboard Cite

Most video advertising campaigns today are still evaluated based on aggregate demographic audience metrics, rather than measures of individual impact or even individual demographic reach. To fit in with advertisers' evaluations, campaigns must be optimized toward validation by third-party measurement companies, which act as "oracles" in assessing ground truth. However, information is only available from such oracles in aggregate, leading to a setting with incomplete ground truth. We explore methods for building probabilistic classification models using these aggregate data. If they perform well, such models can be used to create new "engineered" segments that perform better than existing segments, in terms of lift and/or reach. We focus on the setting where companies already have machinery in place for high-performance predictive modeling from traditional, individual-level data. We show that model building, evaluation, and selection can be reliably carried out even with access only to aggregate ground truth data. We show various concrete results, highlighting confounding aspects of the problem, such as the tendency for pre-existing "in-target" segments actually to comprise biased subpopulations, which has implications both for campaign performance and modeling performance. The paper's main results show that these methods lead to engineered segments that can substantially improve lift and/or reach-as verified by a leading thirdparty oracle. For example, for lifts of 2-3X, segment reach can be increased to 57 times that of comparable, pre-existing segments.

show abstract

Scalable hands-free transfer learning for online advertising

Cited by 33 publications

References 9 publications

Targeted display advertising: the case of preferential attachment

Targeted display advertising: the case of preferential attachment

Display Advertising with Real-Time Bidding (RTB) and Behavioural Targeting

Pleasing the advertising oracle

Contact Info

Product

Resources

About