Causal Discovery Research Articles

Heterogeneous Mediation Analysis for Cox Proportional Hazards Model With Multiple Mediators.

This study proposes a heterogeneous mediation analysis for survival data that accommodates multiple mediators and sparsity of the predictors. We introduce a joint modeling approach that links the mediation regression and proportional hazards models through Bayesian additive regression trees with shared typologies. The shared tree component is motivated by the fact that confounders and effect modifiers on the causal pathways linked by different mediators often overlap. A sparsity-inducing prior is incorporated to capture the most relevant confounders and effect modifiers on different causal pathways. The individual-specific interventional direct and indirect effects are derived on the scale of the logarithm of hazards and survival function. A Bayesian approach with an efficient Markov chain Monte Carlo algorithm is developed to estimate the conditional interventional effects through the Monte Carlo implementation of the mediation formula. Simulation studies are conducted to verify the empirical performance of the proposed method. An application to the ACTG175 study further demonstrates the method's utility in causal discovery and heterogeneity quantification.

How the Spacer Influences the Stability of 2D Perovskites?

Two-dimensionallead halide perovskites (2D HPs) represent as an emerging class of materials given their tunable optoelectronic properties and long-term stability in perovskite solar cells. However, the ever-growing field of optoelectronic devices using 2D HPs requires fundamental understanding of the influence of the spacer on the physiochemical properties and stability of perovskites as well as establish which cation properties are closely related to suppress the halogen ion mobility. This study focuses on investigating the influence of organic spacers with intrinsic properties (e.g., rigidity and flexibility, special groups) and variations of material dimensions on the stability of halogen ions and inorganic frameworks in 2D HPs. It is found that the perovskite structure composed of rigidity molecules owns better stability of halogen ion and inorganic framework than that of flexible molecules. The stability of ions exhibits a negative correlation with the dimensions of perovskite. More importantly, a simple descriptor for measuring the stability of halogen ions in 2D HPs is constructed. By causal discovery algorithms with more physical and chemical significance, the Kappa shape index, number of rotatable bonds, and aromatic carbocycles in organic spacers are identified as causal and important features for the stability of halogen ions in 2D HPs.

Mixed-variable graphical modeling framework towards risk prediction of hospital-acquired pressure injury in spinal cord injury individuals

Developing machine learning (ML) methods for healthcare predictive modeling requires absolute explainability and transparency to build trust and accountability. Graphical models (GM) are key tools for this but face challenges like small sample sizes, mixed variables, and latent confounders. This paper presents a novel learning framework addressing these challenges by integrating latent variables using fast causal inference (FCI), accommodating mixed variables with predictive permutation conditional independence tests (PPCIT), and employing a systematic graphical embedding approach leveraging expert knowledge. This method ensures a transparent model structure and an explainable feature selection and modeling approach, achieving competitive prediction performance. For real-world validation, data of hospital-acquired pressure injuries (HAPI) among individuals with spinal cord injury (SCI) were used, where the approach achieved a balanced accuracy of 0.941 and an AUC of 0.983, outperforming most benchmarks. The PPCIT method also demonstrated superior accuracy and scalability over other benchmarks in causal discovery validation on synthetic datasets that closely resemble our real dataset. This holistic framework effectively addresses the challenges of mixed variables and explainable predictive modeling for disease onset, which is crucial for enabling transparency and interpretability in ML-based healthcare.

Federated Learning of Generalized Linear Causal Networks.

Causal discovery, the inference of causal relations among variables from data, is a fundamental problem of science. Nowadays, due to an increased awareness of data privacy concerns, there has been a shift towards distributed data collection, processing and storage. To meet the pressing need for distributed causal discovery, we propose a novel federated DAG learning method called distributed annealing on regularized likelihood score (DARLS) to learn a causal graph from data stored on multiple clients. DARLS simulates an annealing process to search over the space of topological sorts, where the optimal graphical structure compatible with a sort is found by distributed optimization. This distributed optimization relies on multiple rounds of communication between local clients and a central server to estimate the graphical structure. We establish its convergence to the solution obtained by an oracle with access to all the data. To the best of our knowledge, DARLS is the first distributed method for learning causal graphs with such finite-sample oracle guarantees. To establish the consistency of DARLS, we also derive new identifiability results for causal graphs parameterized by generalized linear models, which could be of independent interest. Through extensive simulation studies and a real-world application, we show that DARLS outperforms existing federated learning methods and is comparable to oracle methods on pooled data, demonstrating its great advantages in estimating causal networks from distributed data.

Causal reasoning in Software Quality Assurance: A systematic review

Context:Software Quality Assurance (SQA) is a fundamental part of software engineering to ensure stakeholders that software products work as expected after release in operation. Machine Learning (ML) has proven to be able to boost SQA activities and contribute to the development of quality software systems. In this context, Causal Reasoning is gaining increasing interest as a methodology to go beyond a purely data-driven approach by exploiting the use of causality for more effective SQA strategies. Objective:Provide a broad and detailed overview of the use of causal reasoning for SQA activities, in order to support researchers to access this research field, identifying room for application, main challenges and research opportunities. Methods:A systematic review of the scientific literature on causal reasoning for SQA. The study has found, classified, and analyzed 86 articles, according to established guidelines for software engineering secondary studies. Results:Results highlight the primary areas within SQA where causal reasoning has been applied, the predominant methodologies used, and the level of maturity of the proposed solutions. Fault localization is the activity where causal reasoning is more exploited, especially in the web services/microservices domain, but other tasks like testing are rapidly gaining popularity. Both causal inference and causal discovery are exploited, with the Pearl’s graphical formulation of causality being preferred, likely due to its intuitiveness. Tools to favour their application are appearing at a fast pace – most of them after 2021. Conclusions:The findings show that causal reasoning is a valuable means for SQA tasks with respect to multiple quality attributes, especially during V&V, evolution and maintenance to ensure reliability, while it is not yet fully exploited for phases like requirements engineering and design. We give a picture of the current landscape, pointing out exciting possibilities for future research.

Deep‐Learning Based Causal Inference: A Feasibility Study Based on Three Years of Tectonic‐Climate Data From Moxa Geodynamic Observatory

AbstractHighly sensitive laser strainmeters at Moxa Geodynamic Observatory (MGO) measure motions of the upper Earth's crust. Since the mountain overburden of the laser strainmeters installed in the gallery of the observatory is relatively low, the recorded time series are strongly influenced by local meteorological phenomena. To estimate the nonlinear effect of the meteorological variables on strain measurements in a non‐stationary environment, advanced methods capable of learning the nonlinearity and discovering causal relationships in the non‐stationary multivariate tectonic‐climate time series are needed. Methods for causal inference generally perform well in identifying linear causal relationships but often struggle to retrieve complex nonlinear causal structures prevalent in real‐world systems. This work presents a novel model invariance‐based causal discovery (CDMI) method that utilizes deep networks to model nonlinearity in a multivariate time series system. We propose to use the theoretically well‐established Knockoffs framework to generate in‐distribution, uncorrelated copies of the original data as interventional variables and test the model invariance for causal discovery. To deal with the non‐stationary behavior of the tectonic‐climate time series recorded at the MGO, we propose a regime identification approach that we apply before causal analysis to generate segments of time series that possess locally consistent statistical properties. First, we evaluate our method on synthetically generated time series by comparing it to other causal analysis methods. We then investigate the hypothesized effect of meteorological variables on strain measurements. Our approach outperforms other causality methods and provides meaningful insights into tectonic‐climate causal interactions.

Causal inference of urban heat island effect and its spatial heterogeneity: A case study of Wuhan, China

The accelerated urbanization process has exacerbated the urban heat island effect, leading to significant negative impacts on the physical and mental health of urban residents. A deeper understanding of the mechanisms underlying the urban heat island phenomenon is essentially beneficial for providing scientific supports towards improving the urban thermal environment. To address the challenge of effectively depicting the complex interactions among urban environment factors, this study employed the Peter-Clark causal discovery algorithm to analyze the causal structure of urban thermal driving factors, and validated the effectiveness of the 6 key factors directly influencing the surface urban heat island intensity (SUHII). In response to the inadequacy of existing big data causal inference tools in assessing the spatial heterogeneity of causal effects, this study proposed a method for evaluating the causal effects on SUHII and their spatial heterogeneity based on local analysis and geospatial causal principle. The result for 4 different intervention scenarios in this study show that there is obvious spatial heterogeneity in the causal effects of different interventions on SUHII in Wuhan, and that increasing greenery and preserving natural environments is an effective way to mitigate the urban heat island (UHI) effect. This approach, provides a new perspective for studying the phenomenon of UHI and insights of potential approaches for mitigating UHI.

TSLiNGAM: DirectLiNGAM Under Heavy Tails

One of the established approaches to causal discovery consists of combining directed acyclic graphs (DAGs) with structural causal models (SCMs) to describe the functional dependencies of effects on their causes. Possible identifiability of SCMs given data depends on assumptions made on the noise variables and the functional classes in the SCM. For instance, in the LiNGAM model, the functional class is restricted to linear functions and the disturbances have to be non-Gaussian. In this work, we propose TSLiNGAM, a new method for identifying the DAG of a causal model based on observational data. TSLiNGAM builds on DirectLiNGAM, a popular algorithm which uses simple OLS regression for identifying causal directions between variables. TSLiNGAM leverages the non-Gaussianity assumption of the error terms in the LiNGAM model to obtain more efficient and robust estimation of the causal structure. TSLiNGAM is justified theoretically and is studied empirically in an extensive simulation study. It performs significantly better on heavy-tailed and skewed data and demonstrates a high small-sample efficiency. In addition, TSLiNGAM also shows better robustness properties as it is more resilient to contamination. Supplementary materials for this article are available online.

InfoFlowNet: A multi-head attention-based self-supervised learning model with surrogate approach for uncovering brain effective connectivity

Deciphering brain network topology can enhance the depth of neuroscientific knowledge and facilitate the development of neural engineering methods. Effective connectivity, which gauges the directional influences among brain regions, is pivotal for these studies. This research introduces InfoFlowNet, a novel self-supervised learning model designed to infer causal relationships in electroencephalogram (EEG) data, aiming at capturing the complex information flow among brain regions. Specifically, InfoFlowNet employs convolution operations and multi-head self-attention mechanisms with a masking feature, which ignores self-connections to focus on inter-regional dynamics. Additionally, a new causal magnitude estimation is introduced by assessing the impact of shuffled surrogate data on signal prediction to quantify causal influence. Experiments using synthetic data and real EEG datasets demonstrate that InfoFlowNet effectively uncovers time-varying causal relationships. Compared with the Granger causality model (GCM) and the temporal causal discovery framework (TCDF), InfoFlowNet shows superior sensitivity in detecting significant causal edges. For instance, in a psychomotor vigilance task, InfoFlowNet identified 8 out of 16 significant causal edges, significantly outperforming GCM and TCDF. Furthermore, InfoFlowNet successfully passes the whiteness test on the model's residuals, and advanced analyses illustrate the advantages of utilizing multi-head attention and masking mechanisms to capture effective connectivity. In sum, InfoFlowNet represents a significant advancement in the analysis of effective connectivity, offering a deeper understanding of brain network dynamics. The model's ability to discern complex causal interactions supports its potential application in broader neuroscientific research and applications.

Species interactions drive continuous assembly of freshwater communities in stochastic environments

Understanding the factors driving the maintenance of long-term biodiversity in changing environments is essential for improving restoration and sustainability strategies in the face of global environmental change. Biodiversity is shaped by both niche and stochastic processes, however the strength of deterministic processes in unpredictable environmental regimes is highly debated. Since communities continuously change over time and space—species persist, disappear or (re)appear—understanding the drivers of species gains and losses from communities should inform us about whether niche or stochastic processes dominate community dynamics. Applying a nonparametric causal discovery approach to a 30-year time series containing annual abundances of benthic invertebrates across 66 locations in New Zealand rivers, we found a strong negative causal relationship between species gains and losses directly driven by predation indicating that niche processes dominate community dynamics. Despite the unpredictable nature of these system, environmental noise was only indirectly related to species gains and losses through altering life history trait distribution. Using a stochastic birth-death framework, we demonstrate that the negative relationship between species gains and losses can not emerge without strong niche processes. Our results showed that even in systems that are dominated by unpredictable environmental variability, species interactions drive continuous community assembly.

KPI-oriented process monitoring based on causal-weighted partial least squares

With the increasing demand for product quality, Key performance indicator (KPI)-oriented process monitoring plays an important role in modern industrial. Partial least squares (PLS)-based methods are widely adopted for KPI-oriented process monitoring, in which the process variable space is decomposed into a principal component subspace that has a strong correlation with KPIs and a residual subspace unrelated to KPIs, and the two spaces are monitored separately. However, if a KPI-unrelated fault occurs in a variable that has no causal relation with any KPI, but has a spurious correlation with some KPIs because of the existence of confounders, PLS based methods may mis-regard the KPI-unrelated fault as a KPI-related fault. This occurs because as a correlation analysis-based method, PLS cannot discriminate whether a variable is a real cause of KPIs or has a spurious correlation with KPIs. Motivated by this, this paper proposes two causal-weighted PLS methods for KPI-oriented process monitoring by combining the LiNGAM-based causal discovery with PLS, which first calculate causal weights based on the modified LiNGAM algorithm and bootstrap strategy, and then employ the causal weights to reweight the weight vectors of PLS to enhance the influence of causal variables in the KPI-related principal subspace and reduce the influence of spurious correlations. Case studies based on a simulated dataset, Tennessee Eastman Process dataset and field data from finishing rolling mill process show that the proposed methods can significantly reduce the false alarm rate for KPI-unrelated faults (i.e., the probability that a KPI-unrelated fault sample is mis-regarded as a KPI-related fault sample) caused by spurious correlation without significantly compromising the fault detection rate of KPI-related faults.

Valid Inference After Causal Discovery

Causal discovery and causal effect estimation are two fundamental tasks in causal inference. While many methods have been developed for each task individually, statistical challenges arise when applying these methods jointly: estimating causal effects after running causal discovery algorithms on the same data leads to “double dipping,” invalidating the coverage guarantees of classical confidence intervals. To this end, we develop tools for valid post-causal-discovery inference. Across empirical studies, we show that a naive combination of causal discovery and subsequent inference algorithms leads to highly inflated miscoverage rates; on the other hand, applying our method provides reliable coverage while achieving more accurate causal discovery than data splitting.

Bayesian networks in modeling leucocyte interplay following brain irradiation: A comprehensive framework

Background and ObjectiveUnderstanding the intricate interactions among leucocyte subpopulations following radiotherapy is crucial for advancing cancer research and immunology. Recently, interest in recent radiotherapy modalities, such as protons, has increased. Herein, we present a framework utilizing Bayesian networks to uncover these complex relationships via an illustrative example of brain irradiation in rodents. MethodsWe utilized data from 96 healthy C57BL/6 adult mice subjected to either X-ray or proton brain irradiation. Leucocyte subpopulations in the blood collected 12 hours after the final irradiated fraction were quantified. We employed Bayesian networks to detect causal interplay between physiological parameters, radiation variables and circulating leucocytes. The causal structure was learned via the use of the Bayesian information criterion as a scored criterion. Parameter estimation was performed to quantify the strength of the identified causal relationships. Cross-validation was used to validate our Bayesian network model's performance. ResultsIn the X-ray model, we discovered previously undisclosed interactions between NK-cells and neutrophils, and between monocytes and T-CD4+ cells. The proton model revealed an interplay involving T-CD4+ cells and neutrophils. Both X-rays and protons led to heightened interactions between T-CD8+ cells and B cells, indicating their significant role in orchestrating immune responses. Additionally, the proton model displayed strengthened interactions between T-CD4+ and T-CD8+ cells, emphasizing a dynamic and coordinated immune response post-irradiation. Cross-validation results demonstrated the robustness of the Bayesian network model in explaining data uncertainty. ConclusionThe use of Bayesian networks as tools for causal structure discovery has revealed novel insights into the dynamics of immune responses to radiation exposure.

Visual Analysis of Multi-outcome Causal Graphs.

We introduce a visual analysis method for multiple causal graphs with different outcome variables, namely, multi-outcome causal graphs. Multi-outcome causal graphs are important in healthcare for understanding multimorbidity and comorbidity. To support the visual analysis, we collaborated with medical experts to devise two comparative visualization techniques at different stages of the analysis process. First, a progressive visualization method is proposed for comparing multiple state-of-the-art causal discovery algorithms. The method can handle mixed-type datasets comprising both continuous and categorical variables and assist in the creation of a fine-tuned causal graph of a single outcome. Second, a comparative graph layout technique and specialized visual encodings are devised for the quick comparison of multiple causal graphs. In our visual analysis approach, analysts start by building individual causal graphs for each outcome variable, and then, multi-outcome causal graphs are generated and visualized with our comparative technique for analyzing differences and commonalities of these causal graphs. Evaluation includes quantitative measurements on benchmark datasets, a case study with a medical expert, and expert user studies with real-world health research data.

A hybrid constrained continuous optimization approach for optimal causal discovery from biological data.

Understanding causal effects is a fundamental goal of science and underpins our ability to make accurate predictions in unseen settings and conditions. While direct experimentation is the gold standard for measuring and validating causal effects, the field of causal graph theory offers a tantalizing alternative: extracting causal insights from observational data. Theoretical analysis has shown that this is indeed possible, given a large dataset and if certain conditions are met. However, biological datasets, frequently, do not meet such requirements but evaluation of causal discovery algorithms is typically performed on synthetic datasets, which they meet all requirements. Thus, real-life datasets are needed, in which the causal truth is reasonably known. In this work we first construct such a large-scale real-life dataset and then we perform on it a comprehensive benchmarking of various causal discovery methods. We find that the PC algorithm is particularly accurate at estimating causal structure, including the causal direction which is critical for biological applicability. However, PC does only produces cause-effect directionality, but not estimates of causal effects. We propose PC-NOTEARS (PCnt), a hybrid solution, which includes the PC output as an additional constraint inside the NOTEARS optimization. This approach combines PC algorithm's strengths in graph structure prediction with the NOTEARS continuous optimization to estimate causal effects accurately. PCnt achieved best aggregate performance across all structural and effect size metrics. https://github.com/zhu-yh1/PC-NOTEARS.

Development of optimised causal AI and its application in retail banking

Causal artificial intelligence (AI) has emerged as a promising approach in machine learning (ML), as it considers not only correlations but also cause-and-effect relationships in data, resulting in more human-like decision making. The pivotal stages within causal AI involve causal discovery and inferencing, each playing a crucial role in extracting meaningful insights from the data. In the realm of causal discovery, various algorithms have been developed to uncover the underlying cause-and-effect structures within data sets. A notable limitation, however, surfaces when attempting to apply these algorithms to data sets characterised by binary variables. This constraint prompts a crucial examination of the current methodologies and calls for innovative solutions that can seamlessly navigate the complexities of binary variable data sets. This paper proposes an optimised causal discovery algorithm that is integrated with the causal inference method based on the estimation of conditional average treatment effects (CATE) scores. The results present the potential of causal AI in terms of incremental impact on the predictive capability of AI/ML models. And the incremental impact is elucidated by comparing conventional propensity-based modelling and causal AI-based modelling by means of a use case in the field of retail banking.

Can algorithms Replace Expert Knowledge for Causal Inference? A Case Study on Novice Use of Causal Discovery.

With growing interest in causal inference and machine learning among epidemiologists, there is increasing discussion of causal discovery algorithms for guiding covariate selection. We present a case study of novice application of causal discovery tools and attempt to validate the results against a well-established causal relationship. As a case study, we attempted causal discovery of relationships relevant to the effect of adherence on mortality in the placebo arm of the Coronary Drug Project (CDP) dataset. We used four algorithms available as existing software implementations and varied several model inputs. We identified 15 adjustment sets from 17 model parameterizations. When applied to a baseline covariate adjustment analysis, these 15 adjustment sets returned effect estimates with similar magnitude and direction of bias as prior published results. When using methods to control for time-varying confounding, there was generally more residual bias than compared to expert-selected adjustment sets. Although causal discovery algorithms can perform on par with expert knowledge, we do not recommend novice use of causal discovery without the input of experts in causal discovery. Expert support is recommended to aid in choosing the algorithm, selecting input parameters, assessing underlying assumptions, and finalizing selection of the adjustment variables.

Tropical and mid-latitude causal drivers of the eastern Mediterranean Etesians during boreal summer

During boreal summer, large scale subsidence and a persistent northerly flow, known as the Etesians, characterize the tropospheric circulation over the eastern Mediterranean. The Etesians bring clear skies and alleviate the impact of heat waves over the region. The intraseasonal variability of the Etesians and subsidence over the eastern Mediterranean has been thought to be influenced by the South Asian monsoon and atmospheric processes over the North Atlantic. Here, we employ causal effect networks and causal maps, obtained by applying the Peter and Clark Momentary Conditional Independence (PCMCI) causal discovery algorithm, to identify causal precursors of Etesians. We find that both wave train activity over the North Atlantic/North American region and convective activity over South Asia associated with the Indian summer monsoon (ISM) are causally related to the Etesians at 3-day time scale. Thus, intraseasonal ISM variability affects the eastern Mediterranean circulation, though its influence is conveyed via a Middle East ridge. On longer weekly time scale, the mid-latitude influence weakens, while the influence of the tropical convective activity via the Middle East ridge remains stable. Moreover, the heat low over the Arabian Peninsula, a feature strongly responsible for the development of the Etesians, is caused by a stronger Middle East ridge and not by North Atlantic wave activity. Finally, we discuss potential implication for circulation changes in the eastern Mediterranean due to anthropogenic global warming.

Causal Discovery Using Weight-Based Conditional Independence Test

Conditional independence (CI) tests play an essential role in causal discovery from observational data, enabling the measurement of independence between two nodes. However, traditional CI tests ignore the imbalanced occurrence probabilities of node values, which may affect the accuracy of determining independence between nodes. To address this problem, we first introduce a new concept of the Node-imbalance phenomenon to describe the imbalance of node values in the Bayesian network data and analyze the influence of the Node-imbalance phenomenon on the traditional CI tests, then we propose a W eight-based C onditional I ndependence test (WCI) to improve the accuracy of CI tests in the presence of Node-imbalance. In the experiments, we verify that WCI effectively measures the dependency between nodes in the Node-imbalance phenomenon compared with the traditional independence tests, and the state-of-the-art causal discovery algorithms reduce the number of false causal orientations through WCI.

Causality-based adversarial attacks for robust GNN modelling with application in fault detection

Fault detection techniques based on graph neural networks have been a trending topic. With the issue of poor robustness, the accuracy relies highly on the quality of the monitoring data. Numerous scholars have come up with robust GNN models. However, the model's accuracy remains low when it comes to solving tasks like graph-level fault detection. In this work, the authors propose several causality-based adversarial attacks that are designed with reference to the principles of causal discovery algorithms for generating causal graph models and associated errors. The attack amplifies all possible types of raw errors present in the data, which allows the trained model to be robust and accurate enough to maintain high error detection accuracy with the proposed adversarial elimination regularization. A real dataset considering high-speed train braking system is considered as case study. Three typical graph neural network models including classical GCN, robust GCN and median GCN are taken as base models to verify the validity of the modelling framework. The results prove that the causality-based adversarial attacks proposed in this work can effectively improve all the base models’ robustness with low-quality monitoring data.