Abstract Learning causal effects from observational data greatly benefits a variety of domains such as health care, education, and sociology. For instance, one could estimate the impact of a new drug on specific individuals to assist clinical planning and improve the survival rate. In this paper, we focus on studying the problem of estimating the Conditional Average Treatment Effect (CATE) from observational data. The challenges for this problem are two-fold: on the one hand, we have to derive a causal estimator to estimate the causal quantity from observational data, in the presence of confounding bias; on the other hand, we have to deal with the identification of the CATE when the distributions of covariates over the treatment group units and the control units are imbalanced. To overcome these challenges, we propose a neural network framework called Adversarial Balancing-based representation learning for Causal Effect Inference (ABCEI), based on recent advances in representation learning. To ensure the identification of the CATE, ABCEI uses adversarial learning to balance the distributions of covariates in the treatment and the control group in the latent representation space, without any assumptions on the form of the treatment selection/assignment function. In addition, during the representation learning and balancing process, highly predictive information from the original covariate space might be lost. ABCEI can tackle this information loss problem by preserving useful information for predicting causal effects under the regularization of a mutual information estimator. The experimental results show that ABCEI is robust against treatment selection bias, and matches/outperforms the state-of-the-art approaches. Our experiments show promising results on several datasets, encompassing several health care (and other) domains.

Objective: Demonstrate how observational causal inference methods can generate insights into the impact of chronic disease combinations on patients' 30-day hospital readmissions. Materials and Methods: Causal effect estimation was used to quantify the impact of each risk factor scenario (ie, chronic disease combination) associated with chronic kidney disease and heart failure (HF) for adult Medicaid beneficiaries with initial hospitalizations in 2 New York State counties. The experimental protocol: (1) created matched risk factor and comparator groups, (2) assessed covariate balance in the matched groups, and (3) estimated causal effects and their statistical significance. Causality lattices summarized the impact of chronic disease comorbidities on readmissions. Results: Chronic disease combinations were ordered with respect to their causal impact on readmissions. Of disease combinations associated with HF, the combination of HF, coronary artery disease, and tobacco abuse (in that order) had the highest causal effect on readmission rate (+22.3%); of disease combinations associated with chronic kidney disease, the combination of chronic kidney disease, coronary artery disease, and diabetes had the highest effect (+9.5%). Discussion: Multi-hypothesis causal analysis reveals the effects of chronic disease comorbidities on health outcomes. Understanding these effects will guide the development of health care programs that address unique care needs of different patient subpopulations. Additionally, these insights bring new attention to individuals at high risk for readmission based on chronic disease comorbidities, allowing for more personalized attention and prioritization of care. Conclusion: Multi-hypothesis causal analysis, a new methodological tool, generates meaningful insights from health care claims data, guiding the design of care and intervention programs.

This paper presents a new pedagogical paradigm ``Crowdlearning'', where students experience deeper learning through collaboratively creating learning materials for each other. Crowdlearning practice is envisioned to produce large ``banks'' of subject matter problems generated by students themselves, in a crowdsourced way, as the students learn new subjects; these problems can then serve as learning and assessment materials usable at scale. This paper overviews the motivation for the development of Crowdlearning as a teaching practice and the theoretical drivers behind it. The paper then reports on preliminary field studies and experiences suggesting that Crowdlearning has a solid potential for adoption in STEM.

The challenges of assessing the "health" of online social media platforms and strategically growing them are recognized by many practitioners and researchers. For those platforms that primarily rely on user-generated content, the reach -- the degree of participation referring to the percentage and involvement of users -- is a key indicator of success. This paper lays a theoretical foundation for measuring engagement as a driver of reach that achieves growth via positive externality effects. The paper takes a game theoretic approach to quantifying engagement, viewing a platform's social capital as a cooperatively created value and finding a fair distribution of this value among the contributors. It introduces engagement capacity, a measure of the ability of users and user groups to engage peers, and formulates the Engaging Team Formation Problem (EngTFP) to identify the sets of users that "make a platform go". We show how engagement capacity can be useful in characterizing forum user behavior and in the reach maximization efforts. We also stress how engagement analysis differs from influence measurement. Computational investigations with Twitter and Health Forum data reveal the properties of engagement capacity and the utility of EngTFP.

This paper presents an information theory-driven matching methodology for making causal inference from observational data. The paper adopts a “potential outcomes framework” view on evaluating the strength of cause-effect relationships: the population-wide average effects of binary treatments are estimated by comparing two groups of units -- the treated and untreated (control). To reduce the bias in such treatment effect estimation, one has to compose a control group in such a way that across the compared groups of units, treatment is independent of the units' covariates. This requirement gives rise to a subset selection / matching problem. This paper presents the models and algorithms that solve the matching problem by minimizing the mutual information (MI) between the covariates and the treatment variable. Such a formulation becomes tractable thanks to the derived optimality conditions that tackle the non-linearity of the sample-based MI function. Computational experiments with mixed integer-programming formulations and four matching algorithms demonstrate the utility of MI based matching for causal inference studies. The algorithmic developments culminate in a matching heuristic that allows for balancing the compared groups in polynomial (close to linear) time, thus allowing for treatment effect estimation with large data sets.

This paper reports on the utility of social network analysis methods in the data fusion domain. Given fused data that combine multiple intelligence reports from the same environment, social network extraction and high value individual (HVI) identification are of interest. The research on the feasibility of such activities may help not only in methodological developments in network science but also in testing and evaluation of fusion quality. This paper offers a parallel computing-based methodology to extract a social network of individuals from fused data, captured as a cumulative associated data graph (CDG). To obtain the desired social network, two approaches including a hop count weighted and a path salience approach are developed and compared. A supervised learning framework is implemented for parameterizing the extraction algorithms. Parameters utilized in the extraction algorithm consider paths between individuals within the social network, weighing relationships between these individuals based on the count weighted and the path salience calculation methodologies. An overall link strength value is then calculated by aggregating path hop count weights and saliences between unique individual pairs for the hop count weighted and path salience approaches, respectively. Ordered centrality-based HVI lists are obtained from the CDGs constructed from the Sunni criminal thread and Bath'est resurgence threads of the SYNCOIN data set, under various fusion system settings. The reported results shed light on the sensitivity of betweenness, closeness, and degree centrality metrics to fused graph inputs and the role of HVI identification as a test and evaluation tool for fusion process optimization. The computational results demonstrate superiority of path salience approach in identifying HVIs. The insights generated by these approaches and directions for future research are discussed.

Online pro-health social networks have grown in the past decade. This position paper argues for the need of prescriptive research that would enable models of online interventions capable of assessing the effects imposed on social networks from the outside, e.g., by a strategic decision-maker (supervisor) possessing the full knowledge of the network dynamics. The paper explains the importance of explicit modeling of weak, acquaintance-type ties for this purpose.

This paper describes traces of user activity around a alleged online social network profile of a Boston Marathon bombing suspect, after the tragedy occurred. The analyzed data, collected with the help of an automatic social media monitoring software, includes the perpetrator's page saved at the time the bombing suspects' names were made public, and the subsequently appearing comments left on that page by other users. The analyses suggest that a timely protection of online media records of a criminal could help prevent a large-scale public spread of communication exchange pertaining to the suspects/criminals' ideas, messages, and connections.

Today's social networks are getting larger, and the need to analyze datasets with millions of nodes and billions of edges is not uncommon any more. As a network of social relationships evolves by the addition of new nodes and edges, fast algorithms are desirable for the recomputation of key network measures such as actor centrality. The distributed computing paradigm offers a scalable approach to addressing the recomputation challenge. This paper develops a Map-Reduce implementation of an incremental All-Pairs Shortest Path (APSP) algorithm. The incremental nature of the approach allows for performing minimal work in updating centrality measures, while the Map-Reduce implementation makes it scalable to large data. The key idea of the incremental APSP algorithm [1] is based on the efficient use of past information about the shortest paths between any node and the neighbors of the newly added node. A presented parallelized version of the algorithm relies on a three-step iterative execution of the "map" and "reduce" jobs. Experiences with its implementation are reported in application to a real-world dataset containing 7115 nodes. The experimental runs were performed using the Amazon's EMR service.

Passenger screening is a critical component of aviation security systems. This paper introduces the multistage sequential passenger screening problem (MSPSP), which models passenger and carry-on baggage screening operations in an aviation security system with the capability of dynamically updating the perceived risk of passengers. The passenger screening operation at an airport terminal is subdivided into multiple screening stages, with decisions made to assign each passenger to one of several available security classes at each such stage. Each passenger's assessed threat value (initially determined by an automated passenger prescreening system) is updated after the passenger proceeds through each screening stage. The objective of MSPSP is to maximize the total security of all passenger screening decisions over a fixed time period, given passenger perceived risk levels and security device performance parameters. An optimal policy for screening passengers in MSPSP is obtained using optimal sequential assignment theory. A Monte Carlo simulation-based heuristic is presented and compared with stochastic sequential assignment and feedback control algorithms. Computational analysis of a two-stage security system provides an assessment of the total security performance.

The use of hydrological (surface wind and sea wave slope) dependences for the calculation of microwave characteristics of the ldquoocean-atmosphererdquo system by Kirchhoff method is discussed here. As an important example, we review results of our experiment on bistatic sensing of the sea surface. In this experiment, the Sun is a powerful natural source of microwave radiation noise for the sea surface sensing. For the receiving of solar radiation in the region of quasi-specular scattering, three two-polarization radiometers (0.8, 2.25, and 6.0 cm) are used. The results of joint hydrological and radiometric experiments (wind speed from 0 to 15 m/s) are reported. As a result, the improved quantitative model for microwave (3-50-GHz band) radiometric calculations in remote sensing of the ldquoocean-atmosphererdquo system is presented.