Learning Decomposed Representations for Treatment Effect Estimation

Wu, Anpeng; Yuan, Junkun; Kuang, Kun; Li, Bo; Wu, Runze; Zhu, Qiang; Zhuang, Yiyu; Wu, Fei

doi:10.1109/tkde.2022.3150807

Cited by 18 publications

(12 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Uplift Modeling. Uplift modeling aims to estimate the incremental effect of taking action on an outcome through causal inference [16,34,39], which has broad applications varying from marketing [8] to medical [1,17,44] domains. Class Transformation methods [2,28]directly estimate the uplift based on two assumptions, i.e., binary outcome variable and balanced dataset between control and treatments, which might not be satisfied in real-world scenarios.…”

Section: Related Workmentioning

confidence: 99%

Intelligent Request Strategy Design in Recommender System

Qian¹,

Xu²,

Lv³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Waterfall Recommender System (RS), a popular form of RS in mobile applications, is a stream of recommended items consisting of successive pages that can be browsed by scrolling. In waterfall RS, when a user finishes browsing a page, the edge (e.g., mobile phones) would send a request to the cloud server to get a new page of recommendations, known as the paging request mechanism. RSs typically put a large number of items into one page to reduce excessive resource consumption from numerous paging requests, which, however, would diminish the RSs' ability to timely renew the recommendations according to users' real-time interest and lead to a poor user experience. Intuitively, inserting additional requests inside pages to update the recommendations with a higher frequency can alleviate the problem. However, previous attempts, including only non-adaptive strategies (e.g., insert requests uniformly), would eventually lead to resource overconsumption. To this end, we envision a new learning task of edge intelligence named Intelligent Request Strategy Design (IRSD). It aims to improve the effectiveness of waterfall RSs by determining the appropriate occasions of request insertion based on users' real-time intention. Moreover, we propose a new paradigm of adaptive request insertion strategy named Uplift-based On-edge Smart Request Framework (AdaRequest). AdaRequest 1) captures the dynamic change of users' intentions by matching their real-time behaviors with their historical interests based on attention-based neural networks. 2) estimates the counterfactual uplift of user purchase brought by an inserted request based on causal inference. 3) determines the final request insertion strategy by maximizing the utility function under online resource constraints. We conduct extensive experiments on both offline dataset and online A/B test to verify the effectiveness of AdaRequest. Remarkably, AdaRequest has been deployed on the Waterfall RS of Taobao and brought over 3% lift on Gross Merchandise Value (GMV).

show abstract

Section: Related Workmentioning

confidence: 99%

Intelligent Request Strategy Design in Recommender System

Qian¹,

Xu²,

Lv³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Representation learning methods [ 1 , 3 , 4 , 5 ] are a kind of important method to deal with the counterfactual problem, which divides the historical data of all the patients into two parts, i.e.,

and

for patients who have received treatment

and

, respectively, with

representing the actual treatment patient

has received and

and

representing the observed survival time and the baseline of patient

, respectively. Then, in representation learning methods, instead of learning

, which encounters the counterfactual problem,

based on historical data

and

based on

are learned separately, and for a new patient, the ITE can be predicted by

[ 3 ].…”

Section: Introductionmentioning

confidence: 99%

“…As mentioned in [ 1 ], the CFR model is also extended to some other improved models, such as those in [ 6 , 7 ]. Additionally, considering the unmeasured confounders, Anpeng Wu et al propose an instrumental variable-based counterfactual regression method, which can also be regarded as an improvement to the CFR model [ 5 ].…”

Section: Introductionmentioning

confidence: 99%

“…This makes it impossible to learn a model for ∆ Y i , i.e., ∆Y i (x i ) = f (x i ), based on historical data, which is useful for predicting the ITE for a new patient before the selecting of treatments. Representation learning methods [1,[3][4][5] are a kind of important method to deal with the counterfactual problem, which divides the historical data of all the patients into two parts, i.e.,…”

Section: Introductionmentioning

confidence: 99%

“…propose an instrumental variable-based counterfactual regression method, which can also be regarded as an improvement to the CFR model [5]. ITE estimation methods based on representation learning (including the CFR method mentioned above) face the censoring problem when they are applied to survival time, because the output y t i =1 i or y t i =0 i denoting survival time in this case will become unavailable (also referred to as censored) if the follow-up of the patient i has been lost or patient i is still alive before the trial ends [4,8,9].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction

Zhao

Zhou

et al. 2022

Entropy

View full text Add to dashboard Cite

The estimation of the Individual Treatment Effect (ITE) on survival time is an important research topic in clinics-based causal inference. Various representation learning methods have been proposed to deal with its three key problems, i.e., reducing selection bias, handling censored survival data, and avoiding balancing non-confounders. However, none of them consider all three problems in a single method. In this study, by combining the Counterfactual Survival Analysis (CSA) model and Dragonnet from the literature, we first propose a CSA–Dragonnet to deal with the three problems simultaneously. Moreover, we found that conclusions from traditional Randomized Controlled Trials (RCTs) or Retrospective Cohort Studies (RCSs) can offer valuable bound information to the counterfactual learning of ITE, which has never been used by existing ITE estimation methods. Hence, we further propose a CSA–Dragonnet with Embedded Prior Knowledge (CDNEPK) by formulating a unified expression of the prior knowledge given by RCTs or RCSs, inserting counterfactual prediction nets into CSA–Dragonnet and defining loss items based on the bounds for the ITE extracted from prior knowledge. Semi-synthetic data experiments showed that CDNEPK has superior performance. Real-world experiments indicated that CDNEPK can offer meaningful treatment advice.

show abstract