• Corpus ID: 246430371

Challenges and approaches to privacy preserving post-click conversion prediction

  title={Challenges and approaches to privacy preserving post-click conversion prediction},
  author={Conor O'Brien and Arvind Thiagarajan and Sourav Das and Rafael Barreto and Chetan Kumar Verma and Tim Hsu and James Neufeld and Jonathan J. Hunt},
Online advertising has typically been more personalized than offline advertising, through the use of machine learning models and real-time auctions for ad targeting. One specific task, predicting the likelihood of conversion (i.e. the probability a user will purchase the advertised product), is crucial to the advertising ecosystem for both targeting and pricing ads. Currently, these models are often trained by observing individual user behavior, but, increasingly, regulatory and technical… 

Figures from this paper


An Analysis Of Entire Space Multi-Task Models For Post-Click Conversion Prediction
An ablation approach is used to systematically study recent approaches that incorporate both multitask learning and “entire space modeling” which train the CVR on all logged examples rather than learning a conditional likelihood of conversion given clicked, and shows that several different approaches result in similar levels of positive transfer.
Optimized Cost per Click in Taobao Display Advertising
A bid optimizing strategy called optimized cost per click (OCPC) is proposed which automatically adjusts the bid to achieve finer matching of bid and traffic quality of page view (PV) request granularity and yields substantially better results than previous fixed bid manner.
Real-time bidding for online advertising: measurement and analysis
An empirical analysis and measurement of a production ad exchange is provided, observing that periodic patterns occur in various statistics including impressions, clicks, bids, and conversion rates, which suggest time-dependent models would be appropriate for capturing the repeated patterns in RTB.
Entire Space Multi-Task Model: An Effective Approach for Estimating Post-Click Conversion Rate
This paper model CVR in a brand-new perspective by making good use of sequential pattern of user actions, i.e., impression -> click -> conversion, which is the first public dataset which contains samples with sequential dependence of click and conversion labels for CVR modeling.
The Algorithmic Foundations of Differential Privacy
The preponderance of this monograph is devoted to fundamental techniques for achieving differential privacy, and application of these techniques in creative combinations, using the query-release problem as an ongoing example.
Differentially Private Federated Learning: A Client Level Perspective
The aim is to hide clients' contributions during training, balancing the trade-off between privacy loss and model performance, and empirical studies suggest that given a sufficiently large number of participating clients, this procedure can maintain client-level differential privacy at only a minor cost in model performance.
Privileged Features Distillation at Taobao Recommendations
By distilling the interacted features that are prohibited during serving for CTR and the post-event features for CVR, this work achieves significant improvements over their strong baselines.
On Learning with Label Proportions
This work answers the fundamental question, when and why LLP is possible, by introducing a general framework, Empirical Proportion Risk Minimization (EPRM), and provides a formal guarantee that the individual labels can indeed be learned in the LLP setting.
Cookieless Monster: Exploring the Ecosystem of Web-Based Device Fingerprinting
By analyzing the code of three popular browser-fingerprinting code providers, it is revealed the techniques that allow websites to track users without the need of client-side identifiers and how fragile the browser ecosystem is against fingerprinting through the use of novel browser-identifying techniques.
k-Anonymity: A Model for Protecting Privacy
  • L. Sweeney
  • Computer Science
    Int. J. Uncertain. Fuzziness Knowl. Based Syst.
  • 2002
The solution provided in this paper includes a formal protection model named k-anonymity and a set of accompanying policies for deployment and examines re-identification attacks that can be realized on releases that adhere to k- anonymity unless accompanying policies are respected.