arXiv subjects

Falco J. Bargagli-Stoffi

Publications and source records attributed to Falco J. Bargagli-Stoffi.

At least 19 recordsLinked to original sources

arXiv2026

From Tokens to Policy: Causal and Interpretable Heterogeneous Treatment Effects Identification

Heterogeneous Treatment Effect (HTE) identification is essential to explain the impact of an intervention and optimize our policies accordingly. Existing approaches trade expressivity for interpretability, but, if some active heterogeneity drivers are unmeasured, methods at both ends of this spectrum allow for spurious HTE characterization with no causal reading. In this work, we focus on controlled experiments and argue that an oracle HTE causal characterization via the latent interactors is now within reach, thanks to (i) more extensive pre-treatment measurements, i.e., multi-modal and multi-view, and (ii) scalable representations with minimal human supervision. We then re-frame HTE identification as a Markov-blanket discovery problem on a sufficient and aligned pre-treatment representation, and introduce Neural EXposure Interaction Search (NEXIS), an iterative procedure with a provable selection guarantee. We deploy NEXIS on two anti-poverty programs in Africa, augmenting each with satellite imagery capturing previously unmeasured environmental effect modifiers, leading to new candidate prescriptive guidelines to optimize the programs' next iterations.

cs.LG↗

arXiv2026

Causal Stability Selection

Identifying covariates that modify treatment effects is a central problem in causal inference. Yet existing data-adaptive procedures do not provide finite-sample control over the expected number of false discoveries, risking spurious findings that fail to replicate. We introduce causal stability selection, an algorithm that combines cross-fitted estimation of conditional average treatment effects with integrated path stability selection. The method accommodates arbitrary treatment effect estimators and arbitrary base selectors, and produces a selection set with an explicit, non-asymptotic bound on the expected number of false positives. Under standard causal identifying assumptions and regularity conditions on the base selector, we prove that the estimated selection probabilities converge to their oracle counterparts at the rate of the underlying treatment effect estimator. This establishes a direct connection between treatment effect estimation and effect modifier discovery. We illustrate the method on simulated data and on real-world data from a randomized oncology trial and an observational study of maternal smoking and infant birthweight.

stat.ME↗

arXiv2026

Assessing the Carbon Emissions and Energy Consumption of U.S. Hyperscale Data Centers

The rapid proliferation of hyperscale data centers (HDCs) in the US, mainly driven by the adoption of artificial intelligence, has raised concerns about this industry's environmental footprint. We compiled facility-level information on 403 US hyperscale data centers operating between May 2024 and April 2025 and estimated their electricity consumption, electricity sources, and attributable CO2 emissions. Across different facility-load scenarios, these HDCs consumed approximately 68-99 TWh of electricity and were associated with about 37-54 million metric tons of CO2. Under the central scenario, HDC electricity demand corresponded to approximately 1.8% of total US electricity consumption, with roughly 54% of attributed generation supplied by fossil-fuel sources. The HDC electricity-weighted average carbon intensity was approximately 545 gCO2/kWh, about 48% above the contemporaneous US national grid-average carbon intensity of 370 gCO2/kWh. Our approach provides an attributional tool for assessing the environmental footprint of hyperscale data centers using the most recent EPA eGRID plant-level data.

cs.AI↗

arXiv2026

Beyond Explainable AI (XAI): An Overdue Paradigm Shift and Post-XAI Research Directions

This study provides a cross-disciplinary examination of Explainable Artificial Intelligence (XAI) approaches-focusing on deep neural networks (DNNs) and large language models (LLMs)-and identifies empirical and conceptual limitations in current XAI. We discuss critical symptoms that stem from deeper root causes (i.e., two paradoxes, two conceptual confusions, and five false assumptions). These fundamental problems within the current XAI research field reveal three insights: experimentally, XAI exhibits significant flaws; conceptually, it is paradoxical; and pragmatically, further attempts to reform the paradoxical XAI might exacerbate its confusion-demanding fundamental shifts and new research directions. To move beyond XAI's limitations, we propose a four-pronged synthesized paradigm shift toward reliable and certified AI development. These four components include: verification-focused Interactive AI (IAI) to establish scientific community protocols for certifying AI system performance rather than attempting post-hoc explanations, AI Epistemology for rigorous scientific foundations, User-Sensible AI to create context-aware systems tailored to specific user communities, and Model-Centered Interpretability for faithful technical analysis-together offering comprehensive post-XAI research directions.

cs.CY↗

arXiv2026

Transporting Predictions via Double Machine Learning: Predicting Partially Unobserved Students' Outcomes

Educational policymakers often lack data on student outcomes where standardized tests were not administered. Machine learning can predict unobserved outcomes in target populations using source population data. However, covariate distribution differences between populations reduce model transportability, potentially decreasing predictive accuracy and introducing bias. We propose using double machine learning for covariate-shift weighted models. First, we estimate overlap scores -- the probability an observation belongs to the source dataset given covariates. Second, balancing weights, defined as density ratios of target-to-source membership probabilities, reweight individual observations' contributions to the loss function in target outcome prediction models. This downweights source observations less similar to the target population, allowing predictions to rely more on observations with greater overlap. Consequently, predictions become more transportable under covariate shift. We illustrate this framework using student standardized financial literacy scores (FLS) data. Using Bayesian Additive Regression Trees (BART), we predict missing FLS. We find minimal predictive performance differences between weighted and unweighted models, suggesting limited covariate shift in our setting. Nonetheless, our approach provides a principled framework for addressing covariate shift and is broadly applicable to predictive modeling in social and health sciences, where source-target population differences are common.

stat.AP↗

arXiv2026

Bayesian Nonparametrics for Principal Stratification with Continuous Post-Treatment Variables

Principal stratification provides a causal inference framework for investigating treatment effects in the presence of a post-treatment variable. Principal strata play a key role in characterizing the treatment effect by identifying groups of units with the same or similar values for the potential post-treatment variable at all treatment levels. The literature has focused mainly on binary post-treatment variables. Few papers considered continuous post-treatment variables. In the presence of a continuous post-treatment, a challenge is how to identify and characterize meaningful coarsening of the latent principal strata that lead to interpretable principal causal effects. This paper introduces the Confounders-Aware SHared atoms BAyesian mixture (CASBAH), a novel approach for principal stratification with binary treatment and continuous post-treatment variables. CASBAH leverages Bayesian nonparametric priors with an innovative hierarchical structure for the potential post-treatment outcomes that overcomes some of the limitations of previous works. Specifically, the novel features of our method allow for (i) identifying coarsened principal strata through a data-adaptive approach and (ii) providing a comprehensive quantification of the uncertainty surrounding stratum membership. Through Monte Carlo simulations, we show that the proposed methodology performs better than existing methods in characterizing the principal strata and estimating principal effects of the treatment. Finally, CASBAH is applied to a case study in which we estimate the causal effects of US national air quality regulations on pollution levels and health outcomes.

stat.ME↗

arXiv2026

Fair Policy Learning under Bipartite Network Interference: Learning Fair and Cost-Effective Environmental Policies

Numerous studies have shown the harmful effects of airborne pollutants on human health. Vulnerable groups and communities often bear a disproportionately larger health burden due to exposure to airborne pollutants. Thus, there is a need to design policies that effectively reduce the public health burdens while ensuring cost-effective policy interventions. Designing policies that optimally benefit the population while ensuring equity between groups under cost constraints is a challenging statistical and causal inference problem. In the context of environmental policy this is further complicated by the fact that interventions target emission sources but health impacts occur in potentially distant communities due to atmospheric pollutant transport -- a setting known as bipartite network interference (BNI). To address these issues, we propose a fair policy learning approach under BNI. Our approach allows to learn cost-effective policies under fairness constraints even accounting for complex BNI data structures. We derive asymptotic properties and demonstrate finite sample performance via Monte Carlo simulations. Finally, we apply the proposed method to a real-world dataset linking power plant scrubber installations to Medicare health records for more than 2 million individuals in the U.S. Our method determine fair scrubber allocations to reduce mortality under fairness and cost constraints.

stat.ME↗

arXiv2025

Characterizing the Effects of Environmental Exposures on Social Mobility: Bayesian Semi-parametrics for Principal Stratification

Understanding the causal effects of air pollution exposures on social mobility is attracting increasing attention. At the same time, education is widely recognized as a key driver of social mobility. However, the causal pathways linking fine particulate matter (PM2.5) exposure, educational attainment, and social mobility remain largely unexplored. To address this, we adopt the principal stratification approach, which rigorously defines causal effects when a post-treatment variable--educational attainment--is affected by exposure--PM2.5--and may, in turn, affect the primary outcome--social mobility. To estimate the causal effects, we propose a Bayesian semi-parametric method leveraging infinite mixtures for modeling the primary outcome. The proposed method (i) allows flexible modeling of the distribution of the primary potential outcomes, (ii) improves the accuracy of counterfactual imputation--a fundamental problem in causal inference framework--, and (iii) enables the characterization of treatment effects across different values of the post-treatment variable. We evaluate the performance of the proposed methodology through a Monte Carlo simulation study, demonstrating its advantages over existing approaches. Finally, we apply our method to a national dataset of 3,009 counties in the United States to estimate the causal effect of PM2.5 on social mobility, taking into account educational attainment as a post-treatment variable. Our findings indicate that in counties where higher PM2.5 exposure significantly reduces educational attainment social mobility decreases by approximately 5% compared to counties with lower PM2.5 exposure. We also find that in counties where exposure to PM2.5 does not affect educational attainment, social mobility is reduced by approximately 2% hinting at the possibility of further, yet unexplored, pathways connecting air pollution and social mobility.

stat.ME↗

arXiv2025

Towards Optimal Environmental Policies: Policy Learning under Arbitrary Bipartite Network Interference

The substantial effect of air pollution on cardiovascular disease and mortality burdens is well-established. Emissions-reducing interventions on coal-fired power plants -- a major source of hazardous air pollution -- have proven to be an effective, but costly, strategy for reducing pollution-related health burdens. Targeting the power plants that achieve maximum health benefits while satisfying realistic cost constraints is challenging. The primary difficulty lies in quantifying the health benefits of intervening at particular plants. This is further complicated because interventions are applied on power plants, while health impacts occur in potentially distant communities, a setting known as bipartite network interference (BNI). In this paper, we introduce novel policy learning methods based on Q- and A-Learning to determine the optimal policy under arbitrary BNI. We derive asymptotic properties and demonstrate finite sample efficacy in simulations. We apply our novel methods to a comprehensive dataset of Medicare claims, power plant data, and pollution transport networks. Our goal is to determine the optimal strategy for installing power plant scrubbers to minimize ischemic heart disease (IHD) hospitalizations under various cost constraints. We find that annual IHD hospitalization rates could be reduced in a range from 23.37-55.30 per 10,000 person-years through optimal policies under different cost constraints.

cs.LG↗

arXiv2024

Environmental Burden of United States Data Centers in the Artificial Intelligence Era

The rapid proliferation of data centers in the US - driven partly by the adoption of artificial intelligence - has set off alarm bells about the industry's environmental impact. We compiled detailed information on 2,132 US data centers operating between September 2023 and August 2024 and determined their electricity consumption, electricity sources, and attributable CO$_{2}$e emissions. Our findings reveal that data centers accounted for more than 4% of total US electricity consumption - with 56% derived from fossil fuels - generating more than 105 million tons of CO$_{2}$e (2.18% of US emissions in 2023). Data centers' carbon intensity - the amount of CO$_{2}$e emitted per unit of electricity consumed - exceeded the US average by 48%. Our data pipeline and visualization tools can be used to assess current and future environmental impacts of data centers.

cs.CY↗

arXiv2024

Towards a Health-Based Power Grid Optimization in the Artificial Intelligence Era

The electric power sector is one of the largest contributors to greenhouse gas emissions in the world. In recent years, there has been an unprecedented increase in electricity demand driven by the so-called Artificial Intelligence (AI) revolution. Although AI has and will continue to have a transformative impact, its environmental and health impacts are often overlooked. The standard approach to power grid optimization aims to minimize CO$_2$ emissions. In this paper, we propose a new holistic paradigm. Our proposed optimization directly targets the minimization of adverse health outcomes under energy efficiency and emission constraints. We show the first example of an optimal fuel mix allocation problem aiming to minimize the average number of adverse health effects resulting from exposure to hazardous air pollutants with constraints on the average and marginal emissions. We argue that this new health-based power grid optimization is essential to promote truly sustainable technological advances that align both with global climate goals and public health priorities.

eess.SP↗

arXiv2024

Causal Rule Ensemble: Interpretable Discovery and Inference of Heterogeneous Treatment Effects

In health and social sciences, it is critically important to identify subgroups of the study population where there is notable heterogeneity of treatment effects (HTE) with respect to the population average. Decision trees have been proposed and commonly adopted for the data-driven discovery of HTE due to their high level of interpretability. However, single-tree discovery of HTE can be unstable and oversimplified. This paper introduces the Causal Rule Ensemble (CRE), a new method for HTE discovery and estimation using an ensemble-of-trees approach. CRE offers several key features, including 1) an interpretable representation of the HTE; 2) the ability to explore complex heterogeneity patterns; and 3) high stability in subgroups discovery. The discovered subgroups are defined in terms of interpretable decision rules. Estimation of subgroup-specific causal effects is performed via a two-stage approach, for which we provide theoretical guarantees. Through simulations, we show that the CRE method is highly competitive compared to state-of-the-art techniques. Finally, we apply CRE to discover the heterogeneous health effects of exposure to air pollution on mortality for 35.3 million Medicare beneficiaries across the contiguous U.S.

stat.ME↗

arXiv2024

Difference-in-Differences under Bipartite Network Interference: A Framework for Quasi-Experimental Assessment of the Effects of Environmental Policies on Health

Pollution from coal-fired power plants has been linked to substantial health and mortality burdens in the US. In recent decades, federal regulatory policies have spurred efforts to curb emissions through various actions, such as the installation of emissions control technologies on power plants. However, assessing the health impacts of these measures, particularly over longer periods of time, is complicated by several factors. First, the units that potentially receive the intervention (power plants) are disjoint from those on which outcomes are measured (communities), and second, pollution emitted from power plants disperses and affects geographically far-reaching areas. This creates a methodological challenge known as bipartite network interference (BNI). To our knowledge, no methods have been developed for conducting quasi-experimental studies with panel data in the BNI setting. In this study, motivated by the need for robust estimates of the total health impacts of power plant emissions control technologies in recent decades, we introduce a novel causal inference framework for difference-in-differences analysis under BNI with staggered treatment adoption. We explain the unique methodological challenges that arise in this setting and propose a solution via a data reconfiguration and mapping strategy. The proposed approach is advantageous because analysis is conducted at the intervention unit level, avoiding the need to arbitrarily define treatment status at the outcome unit level, but it permits interpretation of results at the more policy-relevant outcome unit level. Using this interference-aware approach, we investigate the impacts of installation of flue gas desulfurization scrubbers on coal-fired power plants on coronary heart disease hospitalizations among older Americans over the period 2003-2014, finding an overall beneficial effect in mitigating such disease outcomes.

stat.ME↗

arXiv2024

What is essential is visible to the eye: Saliency in primary school ranking and its effect on academic achievements

We propose a new strategy to identify the impact of class rank, exploiting a "visible" primary school rank from teachers' exam grades, and an "invisible" rank from unreported standardized test scores. Leveraging a unique panel dataset on Italian students, we show that the visible rank has a substantial impact on students' perceptions, which affects subsequent academic performance. However, the effect of being surrounded by higher-SES or higher-achieving peers remains positive even accounting for the decrease in rank. Higher-ranked students self-select into high schools with higher average student achievements. Finally, exploiting an extensive survey, we identify psychological mechanisms channeling the rank effect.

econ.GN↗

arXiv2023

Heterogeneous Treatment and Spillover Effects under Clustered Network Interference

The bulk of causal inference studies rule out the presence of interference between units. However, in many real-world scenarios, units are interconnected by social, physical, or virtual ties, and the effect of the treatment can spill from one unit to other connected individuals in the network. In this paper, we develop a machine learning method that uses tree-based algorithms and a Horvitz-Thompson estimator to assess the heterogeneity of treatment and spillover effects with respect to individual, neighborhood, and network characteristics in the context of clustered networks and neighborhood interference within clusters. The proposed Network Causal Tree (NCT) algorithm has several advantages. First, it allows the investigation of the treatment effect heterogeneity, avoiding potential bias due to the presence of interference. Second, understanding the heterogeneity of both treatment and spillover effects can guide policy-makers in scaling up interventions, designing targeting strategies, and increasing cost-effectiveness. We investigate the performance of our NCT method using a Monte Carlo simulation study, and we illustrate its application to assess the heterogeneous effects of information sessions on the uptake of a new weather insurance policy in rural China.

stat.ME↗

arXiv2023

Confounder-Dependent Bayesian Mixture Model: Characterizing Heterogeneity of Causal Effects in Air Pollution Epidemiology

Several epidemiological studies have provided evidence that long-term exposure to fine particulate matter (PM2.5) increases mortality risk. Furthermore, some population characteristics (e.g., age, race, and socioeconomic status) might play a crucial role in understanding vulnerability to air pollution. To inform policy, it is necessary to identify groups of the population that are more or less vulnerable to air pollution. In causal inference literature, the Group Average Treatment Effect (GATE) is a distinctive facet of the conditional average treatment effect. This widely employed metric serves to characterize the heterogeneity of a treatment effect based on some population characteristics. In this work, we introduce a novel Confounder-Dependent Bayesian Mixture Model (CDBMM) to characterize causal effect heterogeneity. More specifically, our method leverages the flexibility of the dependent Dirichlet process to model the distribution of the potential outcomes conditionally to the covariates and the treatment levels, thus enabling us to: (i) identify heterogeneous and mutually exclusive population groups defined by similar GATEs in a data-driven way, and (ii) estimate and characterize the causal effects within each of the identified groups. Through simulations, we demonstrate the effectiveness of our method in uncovering key insights about treatment effects heterogeneity. We apply our method to claims data from Medicare enrollees in Texas. We found six mutually exclusive groups where the causal effects of PM2.5 on mortality are heterogeneous.

stat.ME↗

arXiv2023

Machine Learning for Zombie Hunting: Predicting Distress from Firms' Accounts and Missing Values

In this contribution, we propose machine learning techniques to predict zombie firms. First, we derive the risk of failure by training and testing our algorithms on disclosed financial information and non-random missing values of 304,906 firms active in Italy from 2008 to 2017. Then, we spot the highest financial distress conditional on predictions that lies above a threshold for which a combination of false positive rate (false prediction of firm failure) and false negative rate (false prediction of active firms) is minimized. Therefore, we identify zombies as firms that persist in a state of financial distress, i.e., their forecasts fall into the risk category above the threshold for at least three consecutive years. For our purpose, we implement a gradient boosting algorithm (XGBoost) that exploits information about missing values. The inclusion of missing values in our predictive model is crucial because patterns of undisclosed accounts are correlated with firm failure. Finally, we show that our preferred machine learning algorithm outperforms (i) proxy models such as Z-scores and the Distance-to-Default, (ii) traditional econometric methods, and (iii) other widely used machine learning techniques. We provide evidence that zombies are on average less productive and smaller, and that they tend to increase in times of crisis. Finally, we argue that our application can help financial institutions and public authorities design evidence-based policies-e.g., optimal bankruptcy laws and information disclosure policies.

econ.EM↗

arXiv2021

Heterogeneous causal effects with imperfect compliance: a Bayesian machine learning approach

This paper introduces an innovative Bayesian machine learning algorithm to draw interpretable inference on heterogeneous causal effects in the presence of imperfect compliance (e.g., under an irregular assignment mechanism). We show, through Monte Carlo simulations, that the proposed Bayesian Causal Forest with Instrumental Variable (BCF-IV) methodology outperforms other machine learning techniques tailored for causal inference in discovering and estimating the heterogeneous causal effects while controlling for the familywise error rate (or - less stringently - for the false discovery rate) at leaves' level. BCF-IV sheds a light on the heterogeneity of causal effects in instrumental variable scenarios and, in turn, provides the policy-makers with a relevant tool for targeted policies. Its empirical application evaluates the effects of additional funding on students' performances. The results indicate that BCF-IV could be used to enhance the effectiveness of school funding on students' performance.

stat.ME↗