{"title":"Peeking at A/B Tests: Why it matters, and what to do about it","authors":"Ramesh Johari, P. Koomen, L. Pekelis, David Walsh","doi":"10.1145/3097983.3097992","DOIUrl":"https://doi.org/10.1145/3097983.3097992","url":null,"abstract":"This paper reports on novel statistical methodology, which has been deployed by the commercial A/B testing platform Optimizely to communicate experimental results to their customers. Our methodology addresses the issue that traditional p-values and confidence intervals give unreliable inference. This is because users of A/B testing software are known to continuously monitor these measures as the experiment is running. We provide always valid p-values and confidence intervals that are provably robust to this effect. Not only does this make it safe for a user to continuously monitor, but it empowers her to detect true effects more efficiently. This paper provides simulations and numerical studies on Optimizely's data, demonstrating an improvement in detection performance over traditional methods.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133932524","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yaniv Ovadia, Yoni Halpern, Dilip Krishnan, Josh Livni, Daniel E. Newburger, R. Poplin, Tiantian Zha, D. Sculley
{"title":"Learning to Count Mosquitoes for the Sterile Insect Technique","authors":"Yaniv Ovadia, Yoni Halpern, Dilip Krishnan, Josh Livni, Daniel E. Newburger, R. Poplin, Tiantian Zha, D. Sculley","doi":"10.1145/3097983.3098204","DOIUrl":"https://doi.org/10.1145/3097983.3098204","url":null,"abstract":"Mosquito-borne illnesses such as dengue, chikungunya, and Zika are major global health problems, which are not yet addressable with vaccines and must be countered by reducing mosquito populations. The Sterile Insect Technique (SIT) is a promising alternative to pesticides; however, effective SIT relies on minimal releases of female insects. This paper describes a multi-objective convolutional neural net to significantly streamline the process of counting male and female mosquitoes released from a SIT factory and provides a statistical basis for verifying strict contamination rate limits from these counts despite measurement noise. These results are a promising indication that such methods may dramatically reduce the cost of effective SIT methods in practice.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132771494","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A Data Mining Framework for Valuing Large Portfolios of Variable Annuities","authors":"Guojun Gan, Xiangji Huang","doi":"10.1145/3097983.3098013","DOIUrl":"https://doi.org/10.1145/3097983.3098013","url":null,"abstract":"A variable annuity is a tax-deferred retirement vehicle created to address concerns that many people have about outliving their assets. In the past decade, the rapid growth of variable annuities has posed great challenges to insurance companies especially when it comes to valuing the complex guarantees embedded in these products. In this paper, we propose a novel data mining framework to address the computational issue associated with the valuation of large portfolios of variable annuity contracts. The data mining framework consists of two major components: a data clustering algorithm which is used to select representative variable annuity contracts, and a regression model which is used to predict quantities of interest for the whole portfolio based on the representative contracts. A series of numerical experiments are conducted on a portfolio of synthetic variable annuity contracts to demonstrate the performance of our proposed data mining framework in terms of accuracy and speed. The experimental results show that our proposed framework is able to produce accurate estimates of various quantities of interest and can reduce the runtime significantly.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133000180","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Local Algorithm for User Action Prediction Towards Display Ads","authors":"Hongxia Yang, Yada Zhu, Jingrui He","doi":"10.1145/3097983.3098089","DOIUrl":"https://doi.org/10.1145/3097983.3098089","url":null,"abstract":"User behavior modeling is essential in computational advertisement, which builds users' profiles by tracking their online behaviors and then delivers the relevant ads according to each user's interests and needs. Accurate models will lead to higher targeting accuracy and thus improved advertising performance. Intuitively, similar users tend to have similar behaviors towards the displayed ads (e.g., impression, click, conversion). However, to the best of our knowledge, there is not much previous work that explicitly investigates such similarities of various types of user behaviors, and incorporates them into ad response targeting and prediction, largely due to the prohibitive scale of the problem. To bridge this gap, in this paper, we use bipartite graphs to represent historical user behaviors, which consist of both user nodes and advertiser campaign nodes, as well as edges reflecting various types of user-campaign interactions in the past. Based on this representation, we study random-walk-based local algorithms for user behavior modeling and action prediction, whose computational complexity depends only on the size of the output cluster, rather than the entire graph. Our goal is to improve action prediction by leveraging historical user-user, campaign-campaign, and user-campaign interactions. In particular, we propose the bipartite graphs AdvUserGraph accompanied with the ADNI algorithm. ADNI extends the NIBBLE algorithm to AdvUserGraph, and it is able to find the local cluster consisting of interested users towards a specific advertiser campaign. We also propose two extensions of ADNI with improved efficiencies. The performance of the proposed algorithms is demonstrated on both synthetic data and a world leading Demand Side Platform (DSP), showing that they are able to discriminate extremely rare events in terms of their action propensity.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133233467","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Sparse Compositional Local Metric Learning","authors":"J. S. Amand, Jun Huan","doi":"10.1145/3097983.3098153","DOIUrl":"https://doi.org/10.1145/3097983.3098153","url":null,"abstract":"Mahalanobis distance metric learning becomes an especially challenging problem as the dimension of the feature space p is scaled upwards. The number of parameters to optimize grows with space complexity of order O (p 2), making storage infeasible, interpretability poor, and causing the model to have a high tendency to overfit. Additionally, optimization while maintaining feasibility of the solution becomes prohibitively expensive, requiring a projection onto the positive semi-definite cone after every iteration. In addition to the obvious space and computational challenges, vanilla distance metric learning is unable to model complex and multi-modal trends in the data. Inspired by the recent resurgence of Frank-Wolfe style optimization, we propose a new method for sparse compositional local Mahalanobis distance metric learning. Our proposed technique learns a set of distance metrics which are composed of local and global components. We capture local interactions in the feature space, while ensuring that all metrics share a global component, which may act as a regularizer. We optimize our model using an alternating pairwise Frank-Wolfe style algorithm. This serves a dual purpose, we can control the sparsity of our solution, and altogether avoid any expensive projection operations. Finally, we conduct an empirical evaluation of our method with the current state of the art and present the results on five datasets from varying domains.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125801196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
I. Labutov, Yun Huang, Peter Brusilovsky, Daqing He
{"title":"Semi-Supervised Techniques for Mining Learning Outcomes and Prerequisites","authors":"I. Labutov, Yun Huang, Peter Brusilovsky, Daqing He","doi":"10.1145/3097983.3098187","DOIUrl":"https://doi.org/10.1145/3097983.3098187","url":null,"abstract":"Educational content of today no longer only resides in textbooks and classrooms; more and more learning material is found in a free, accessible form on the Internet. Our long-standing vision is to transform this web of educational content into an adaptive, web-scale \"textbook\", that can guide its readers to most relevant \"pages\" according to their learning goal and current knowledge. In this paper, we address one core, long-standing problem towards this goal: identifying outcome and prerequisite concepts within a piece of educational content (e.g., a tutorial). Specifically, we propose a novel approach that leverages textbooks as a source of distant supervision, but learns a model that can generalize to arbitrary documents (such as those on the web). As such, our model can take advantage of any existing textbook, without requiring expert annotation. At the task of predicting outcome and prerequisite concepts, we demonstrate improvements over a number of baselines on six textbooks, especially in the regime of little to no ground-truth labels available. Finally, we demonstrate the utility of a model learned using our approach at the task of identifying prerequisite documents for adaptive content recommendation --- an important step towards our vision of the \"web as a textbook\".","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128380438","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
B. Kitts, M. Krishnan, I. Yadav, Yongbo Zeng, Garrett Badeau, Andrew Potter, Sergey Tolkachov, Ethan Thornburg, Satyanarayana Reddy Janga
{"title":"Ad Serving with Multiple KPIs","authors":"B. Kitts, M. Krishnan, I. Yadav, Yongbo Zeng, Garrett Badeau, Andrew Potter, Sergey Tolkachov, Ethan Thornburg, Satyanarayana Reddy Janga","doi":"10.1145/3097983.3098085","DOIUrl":"https://doi.org/10.1145/3097983.3098085","url":null,"abstract":"Ad-servers have to satisfy many different targeting criteria, and the combination can often result in no feasible solution. We hypothesize that advertisers may be defining these metrics to create a kind of \"proxy target\". We therefore reformulate the standard ad-serving problem to one where we attempt to get as close as possible to the advertiser's multi-dimensional target inclusive of delivery. We use a simple simulation to illustrate the behavior of this algorithm compared to Constraint and Pacing strategies. The system is then deployed in one of the largest video ad-servers in the United States and we show experimental results from live test ads, as well as 6 months of production performance across hundreds of ads. We find that the live ad-server tests match the simulation, and we report significant gains in multi-KPI performance from using the error minimization strategy.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123275012","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"BDT: Gradient Boosted Decision Tables for High Accuracy and Scoring Efficiency","authors":"Yin Lou, M. Obukhov","doi":"10.1145/3097983.3098175","DOIUrl":"https://doi.org/10.1145/3097983.3098175","url":null,"abstract":"In this paper we present gradient boosted decision tables (BDTs). A d-dimensional decision table is essentially a mapping from a sequence of d boolean tests to a real value in {R}. We propose novel algorithms to fit decision tables. Our thorough empirical study suggests that decision tables are better weak learners in the gradient boosting framework and can improve the accuracy of the boosted ensemble. In addition, we develop an efficient data structure to represent decision tables and propose a novel fast algorithm to improve the scoring efficiency for boosted ensemble of decision tables. Experiments on public classification and regression datasets demonstrate that our method is able to achieve 1.5x to 6x speedups over the boosted regression trees baseline. We complement our experimental evaluation with a bias-variance analysis that explains how different weak models influence the predictive power of the boosted ensemble. Our experiments suggest gradient boosting with randomly backfitted decision tables distinguishes itself as the most accurate method on a number of classification and regression problems. We have deployed a BDT model to LinkedIn news feed system and achieved significant lift on key metrics.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114912860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Martin Saveski, Jean Pouget-Abadie, Guillaume Saint-Jacques, Weitao Duan, Souvik Ghosh, Ya Xu, E. Airoldi
{"title":"Detecting Network Effects: Randomizing Over Randomized Experiments","authors":"Martin Saveski, Jean Pouget-Abadie, Guillaume Saint-Jacques, Weitao Duan, Souvik Ghosh, Ya Xu, E. Airoldi","doi":"10.1145/3097983.3098192","DOIUrl":"https://doi.org/10.1145/3097983.3098192","url":null,"abstract":"Randomized experiments, or A/B tests, are the standard approach for evaluating the causal effects of new product features, i.e., treatments. The validity of these tests rests on the \"stable unit treatment value assumption\" (SUTVA), which implies that the treatment only affects the behavior of treated users, and does not affect the behavior of their connections. Violations of SUTVA, common in features that exhibit network effects, result in inaccurate estimates of the causal effect of treatment. In this paper, we leverage a new experimental design for testing whether SUTVA holds, without making any assumptions on how treatment effects may spill over between the treatment and the control group. To achieve this, we simultaneously run both a completely randomized and a cluster-based randomized experiment, and then we compare the difference of the resulting estimates. We present a statistical test for measuring the significance of this difference and offer theoretical bounds on the Type I error rate. We provide practical guidelines for implementing our methodology on large-scale experimentation platforms. Importantly, the proposed methodology can be applied to settings in which a network is not necessarily observed but, if available, can be used in the analysis. Finally, we deploy this design to LinkedIn's experimentation platform and apply it to two online experiments, highlighting the presence of network effects and bias in standard A/B testing approaches in a real-world setting.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124318782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Saurav Ghosh, Prithwish Chakraborty, B. Lewis, M. Majumder, E. Cohn, J. Brownstein, M. Marathe, Naren Ramakrishnan
{"title":"GELL: Automatic Extraction of Epidemiological Line Lists from Open Sources","authors":"Saurav Ghosh, Prithwish Chakraborty, B. Lewis, M. Majumder, E. Cohn, J. Brownstein, M. Marathe, Naren Ramakrishnan","doi":"10.1145/3097983.3098073","DOIUrl":"https://doi.org/10.1145/3097983.3098073","url":null,"abstract":"Real-time monitoring and responses to emerging public health threats rely on the availability of timely surveillance data. During the early stages of an epidemic, the ready availability of line lists with detailed tabular information about laboratory-confirmed cases can assist epidemiologists in making reliable inferences and forecasts. Such inferences are crucial to understand the epidemiology of a specific disease early enough to stop or control the outbreak. However, construction of such line lists requires considerable human supervision and therefore, difficult to generate in real-time. In this paper, we motivate Guided Epidemiological Line List (GELL), the first tool for building automated line lists (in near real-time) from open source reports of emerging disease outbreaks. Specifically, we focus on deriving epidemiological characteristics of an emerging disease and the affected population from reports of illness. GELL uses distributed vector representations (ala word2vec) to discover a set of indicators for each line list feature. This discovery of indicators is followed by the use of dependency parsing based techniques for final extraction in tabular form. We evaluate the performance of GELL against a human annotated line list provided by HealthMap corresponding to MERS outbreaks in Saudi Arabia. We demonstrate that GELL extracts line list features with increased accuracy compared to a baseline method. We further show how these automatically extracted line list features can be used for making epidemiological inferences, such as inferring demographics and symptoms-to-hospitalization period of affected individuals.","PeriodicalId":314049,"journal":{"name":"Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116838385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}