Causeandeffect relations are one of the most valuable types of knowledge sought after throughout the datadriven sciences since they translate into stable and generalizable explanations as well as efficient and robust decisionmaking capabilities. Inferring these relations from data, however, is a challenging task. Two of the most common barriers to this goal are known as confounding and selection biases. The former stems from the systematic bias introduced during the treat ment assignment, while the latter comes from the systematic bias during the collection of units into the sample. In this paper, we consider the problem of identifiability of causal effects when both confounding and selection biases are simultaneously present. We first investigate the problem of identifiability when all the available data is biased. We prove that the algorithm proposed by [Bareinboim and Tian, 2015] is, in fact, complete, namely, whenever the algorithm returns a failure condition, no identifiability claim about the causal relation can be made by any other method. We then generalize this setting to when, in addition to the biased data, another piece of external data is available, without bias. It may be the case that a subset of the covariates could be measured without bias (e.g., from census). We examine the problem of identifiability when a combination of biased and unbiased data is available. We propose a new algorithm that subsumes the current stateoftheart method based on the backdoor criterion.
more »
« less
Identification of Causal Effects in the Presence of Selection Bias
Causeandeffect relations are one of the most valuable types of knowledge sought after throughout the datadriven sciences since they translate into stable and generalizable explanations as well as efficient and robust decisionmaking capabilities. Inferring these relations from data, however, is a challenging task. Two of the most common barriers to this goal are known as confounding and selection biases. The former stems from the systematic bias introduced during the treatment assignment, while the latter comes from the systematic bias during the collection of units into the sample.
In this paper, we consider the problem of identifiability of causal effects when both confounding and selection biases are simultaneously present. We first investigate the problem of identifiability when all the available data is biased. We prove that the algorithm proposed by [Bareinboim and Tian, 2015] is, in fact, complete, namely, whenever the algorithm returns a failure condition, no identifiability claim about the causal relation can be made by any other method. We then generalize this setting to when, in addition to the biased data, another piece of external data is available, without bias. It may be the case that a subset of the covariates could be measured without bias (e.g., from census). We examine the problem of identifiability when a combination of biased and unbiased data is available. We propose a new algorithm that subsumes the current stateoftheart method based on the backdoor criterion.
more »
« less
 Award ID(s):
 1704352
 NSFPAR ID:
 10098081
 Date Published:
 Journal Name:
 Proceedings of the 33rd AAAI Conference on Artificial Intelligence (AAAI)
 Format(s):
 Medium: X
 Sponsoring Org:
 National Science Foundation
More Like this


Selection and confounding biases are the two most common impediments to the applicability of causal inference methods in largescale settings. We generalize the notion of backdoor adjustment to account for both biases and leverage external data that may be available without selection bias (e.g., data from census). We introduce the notion of adjustment pair and present complete graphical conditions for identifying causal effects by adjustment. We further design an algorithm for listing all admissible adjustment pairs in polynomial delay, which is useful for researchers interested in evaluating certain properties of some admissible pairs but not all (common properties include cost, variance, and feasibility to measure). Finally, we describe a statistical estimation procedure that can be performed once a set is known to be admissible, which entails different challenges in terms of finite samples.more » « less

Selection and confounding biases are the two most common impediments to the applicability of causal inference methods in largescale settings. We generalize the notion of backdoor adjustment to account for both biases and leverage external data that may be available without selection bias (e.g., data from census). We introduce the notion of adjustment pair and present complete graphical conditions for identifying causal effects by adjustment. We further design an algorithm for listing all admissible adjustment pairs in polynomial delay, which is useful for researchers interested in evaluating certain properties of some admissible pairs but not all (common properties include cost, variance, and feasibility to measure). Finally, we describe a statistical estimation procedure that can be performed once a set is known to be admissible, which entails different challenges in terms of finite samples.more » « less

null (Ed.)One fundamental problem in causal inference is to learn the individual treatment effects (ITE)  assessing the causal effects of a certain treatment (e.g., prescription of medicine) on an important outcome (e.g., cure of a disease) for each data instance, but the effectiveness of most existing methods is often limited due to the existence of hidden confounders. Recent studies have shown that the auxiliary relational information among data can be utilized to mitigate the confounding bias. However, these works assume that the observational data and the relations among them are static, while in reality, both of them will continuously evolve over time and we refer such data as timeevolving networked observational data. In this paper, we make an initial investigation of ITE estimation on such data. The problem remains difficult due to the following challenges: (1) modeling the evolution patterns of timeevolving networked observational data; (2) controlling the hidden confounders with current data and historical information; (3) alleviating the discrepancy between the control group and the treated group. To tackle these challenges, we propose a novel ITE estimation framework Dynamic Networked Observational Data Deconfounder (\mymodel) which aims to learn representations of hidden confounders over time by leveraging both current networked observational data and historical information. Additionally, a novel adversarial learning based representation balancing method is incorporated toward unbiased ITE estimation. Extensive experiments validate the superiority of our framework when measured against stateoftheart baselines. The implementation can be accessed in https://github.com/jma712/DNDC https://github.com/jma712/DNDC.more » « less

We present CausalSim, a causal framework for unbiased tracedriven simulation. Current tracedriven simulators assume that the interventions being simulated (e.g., a new algorithm) would not affect the validity of the traces. However, realworld traces are often biased by the choices algorithms make during trace collection, and hence replaying traces under an intervention may lead to incorrect results. CausalSim addresses this challenge by learning a causal model of the system dynamics and latent factors capturing the underlying system conditions during trace collection. It learns these models using an initial randomized control trial (RCT) under a fixed set of algorithms, and then applies them to remove biases from trace data when simulating new algorithms. Key to CausalSim is mapping unbiased tracedriven simulation to a tensor completion problem with extremely sparse observations. By exploiting a basic distributional invariance property present in RCT data, CausalSim enables a novel tensor completion method despite the sparsity of observations. Our extensive evaluation of CausalSim on both real and synthetic datasets, including more than ten months of real data from the Puffer video streaming system shows it improves simulation accuracy, reducing errors by 53% and 61% on average compared to expertdesigned and supervised learning baselines. Moreover, CausalSim provides markedly different insights about ABR algorithms compared to the biased baseline simulator, which we validate with a real deploymentmore » « less