Practical Lessons from Predicting Clicks on Ads at Facebook

He, Xinran; Pan, Jeffrey Junfeng; Jin, Ou; Xu, Tian-Bing; Liu, Bo; Xu, Tao; Shi, Yanxin; Atallah, A.; Herbrich, Ralf; Bowers, Stuart; Candela, Joaquin Quiñonero

doi:10.1145/2648584.2648589

Cited by 729 publications

(453 citation statements)

References 8 publications

Supporting

Mentioning

428

Contrasting

Unclassified

Order By: Relevance

“…Typically, the response prediction problem is formulated as a regression problem with prediction likelihood as the training objective [23,9,1,21]. From the methodology view, linear models such as logistic regression [14] and non-linear models such as tree-based model [10] and factorization machines [19,21] are commonly used. Other variants include Bayesian probit regression [9], FTRFL [24] in factorization machine, and convolutional neural network learning framework [17].…”

Section: Related Workmentioning

confidence: 99%

“…Thus similar to [10], the negative down-sampling and the corresponding calibration methods are adopted in the experiment. The online A/B test is conducted on an operational real-time bidding platform run by YOYI.…”

Section: Datasetsmentioning

confidence: 99%

“…CE -The logistic regression model [10,18] is widely used in many DSP platforms to make predictions of user feedback. This model takes cross entropy as its optimization objective and has the gradient as Eq.…”

Section: Compared Settingsmentioning

confidence: 99%

“…First, the CTR/CVR 1 estimation is formulated as a binary regression problem, which can be solved by such as logistic regression with SGD [23] or FTRL learning [18], Bayesian probit regression [9], gradient boosting regression trees [10], and Factorization Machines [21]. The common objective in this stage is to make the estimation as accurate as possible by, for instance, minimizing the error between the predictions and ground truth user responses.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

User Response Learning for Directly Optimizing Campaign Performance in Display Advertising

Ren

Zhang

Rong

et al. 2016

Proceedings of the 25th ACM International on Conference on Information and Knowledge Management

View full text Add to dashboard Cite

Learning and predicting user responses, such as clicks and conversions, are crucial for many Internet-based businesses including web search, e-commerce, and online advertising. Typically, a user response model is established by optimizing the prediction accuracy, e.g., minimizing the error between the prediction and the ground truth user response. However, in many practical cases, predicting user responses is only part of a rather larger predictive or optimization task, where on one hand, the accuracy of a user response prediction determines the final (expected) utility to be optimized, but on the other hand, its learning may also be influenced from the follow-up stochastic process. It is, thus, of great interest to optimize the entire process as a whole rather than treat them independently or sequentially. In this paper, we take real-time display advertising as an example, where the predicted user's ad click-through rate (CTR) is employed to calculate a bid for an ad impression in the second price auction. We reformulate a common logistic regression CTR model by putting it back into its subsequent bidding context: rather than minimizing the prediction error, the model parameters are learned directly by optimizing campaign profit. The gradient update resulted from our formulations naturally fine-tunes the cases where the market competition is high, leading to a more costeffective bidding. Our experiments demonstrate that, while maintaining comparable CTR prediction accuracy, our proposed user response learning leads to campaign profit gains as much as 78.2% for offline test and 25.5% for online A/B test over strong baselines.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Datasetsmentioning

confidence: 99%

Section: Compared Settingsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

User Response Learning for Directly Optimizing Campaign Performance in Display Advertising

Ren

Zhang

Rong

et al. 2016

Proceedings of the 25th ACM International on Conference on Information and Knowledge Management

View full text Add to dashboard Cite

show abstract

“…Other examples in science include astronomy [3] and bioinformatics [4]. Industry is clearly leading the way in the field, especially big companies like Google, Amazon or Facebook, which mine customers' data for sales and marketing purposes [5]. Smaller-size organisations also have the means to collect and analyse fairly big amounts of data, mainly thanks to open source tools like Hadoop [6].…”

Section: Introductionmentioning

confidence: 99%

SWAN: A service for interactive analysis in the cloud

Piparo

Tejedor

Mató

et al. 2018

Future Generation Computer Systems

View full text Add to dashboard Cite

SWAN (Service for Web based ANalysis) is a platform to perform interactive data analysis in the cloud. SWAN allows users to write and run their data analyses with only a web browser, leveraging on the widely-adopted Jupyter notebook interface. The user code, executions and data live entirely in the cloud. SWAN makes it easier to produce and share results and scientific code, access scientific software, produce tutorials and demonstrations as well as preserve analyses. Furthermore, it is also a powerful tool for non-scientific data analytics.This paper describes how a pilot of the SWAN service was implemented and deployed at CERN. Its backend combines state-of-the-art software technologies with a set of existing IT services such as user authentication, virtual computing infrastructure, mass storage, file synchronisation and sharing, specialised clusters and batch systems.The added value of this combination of services is discussed, with special focus on the opportunities offered by the CERNBox service and its massive storage backend, EOS. In particular, it is described how a cloud-based analysis model benefits from synchronised storage and sharing capabilities.

show abstract

Estimating rates of rare events through a multidimensional dynamic hierarchical Bayesian framework

Yang

Ormándi

Tsao

et al. 2016

Appl Stoch Models Bus & Ind

View full text Add to dashboard Cite

We consider the problem of estimating occurrence rates of rare events for extremely sparse data using pre-existing hierarchies and selected features to perform inference along multiple dimensions. In particular, we focus on the problem of estimating click rates for {Advertiser, Publisher, and User} tuples where both the Advertisers and the Publishers are organized as hierarchies that capture broad contextual information at different levels of granularities. Typically, the click rates are low, and the coverage of the hierarchies and dimensions is sparse. To overcome these difficulties, we decompose the joint prior of the three-dimensional click-through rate using tensor decomposition and propose a multidimensional hierarchical Bayesian framework (abbreviated as MadHab). We set up a specific framework of each dimension to model dimension-specific characteristics. More specifically, we consider the hierarchical beta process prior for the Advertiser dimension and for the Publisher dimension respectively and a feature-dependent mixture model for the User dimension. Besides the centralized implementation, we propose two distributed algorithms through MapReduce and Spark for inferences, which make the model highly scalable and suited for large scale data mining applications. We demonstrate that on a real world ads campaign platform, our framework can effectively discriminate extremely rare events in terms of their click propensity.

show abstract

Practical Lessons from Predicting Clicks on Ads at Facebook

Cited by 729 publications

References 8 publications

User Response Learning for Directly Optimizing Campaign Performance in Display Advertising

User Response Learning for Directly Optimizing Campaign Performance in Display Advertising

SWAN: A service for interactive analysis in the cloud

Estimating rates of rare events through a multidimensional dynamic hierarchical Bayesian framework

Contact Info

Product

Resources

About