Article
Mathematical & Computational Biology
Siyun Yang, Elizabeth Lorenzi, Georgia Papadogeorgou, Daniel M. Wojdyla, Fan Li, Laine E. Thomas
Summary: This article introduces analytical methods and visualization tools for causal subgroup analysis, including subgroup weighted average treatment effect and overlap weighting method. The proposed methods aim to achieve balance within subgroups and to address the bias-variance tradeoff in SGA. The Connect-S plot is designed for visualizing subgroup covariate balance.
STATISTICS IN MEDICINE
(2021)
Article
Engineering, Industrial
Wenli Liu, Fenghua Liu, Weili Fang, Peter E. D. Love
Summary: This paper addresses the issue of interpretability and transparency in machine learning models used for evaluating safety risks in tunnel construction. By utilizing the concept of 'eXplainable AI' (XAI) and causal discovery and reasoning, the authors develop a method to analyze and interpret geotechnical risks in tunnel construction. The proposed approach includes a sparse nonparametric and nonlinear directed acyclic diagram (DAG), a multiple linear regression model, and a probability-based reasoning model. The feasibility and effectiveness of the approach are validated through a case study on a tunnel project in Wuhan, China, showing accurate explanation of data-driven risk assessment results.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Computer Science, Theory & Methods
Pingchuan Ma, Zhenlan Ji, Qi Pang, Shuai Wang
Summary: This paper proposes a new approach to enforce privacy for causal discovery algorithms based on functional causal models. It introduces a differentially private causal discovery algorithm called NoLeaks, along with a highly efficient numerical optimization algorithm. Evaluation results show that NoLeaks achieves comparable or even superior performance compared to the state-of-the-art approaches, and it can scale smoothly to large datasets.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2022)
Article
Computer Science, Information Systems
Boxiang Zhao, Shuliang Wang, Lianhua Chi, Qi Li, Xiaojia Liu, Jing Geng
Summary: This study proposes a new graph structure called Causal Star Graph (CSG) and a corresponding framework called Causal Discovery via Causal Star Graphs (CD-CSG) to address the limitations of existing causal discovery methods. By conducting generalized learning in CSGs, the causal directions in directed acyclic graphs can be accurately identified. Experimental results show that CD-CSG can effectively identify causal relationships between variables and outperforms existing models in terms of accuracy.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2023)
Article
Computer Science, Artificial Intelligence
Hugo M. Proenca, Peter Grunwald, Thomas Back, Matthijs van Leeuwen
Summary: This paper introduces the problem of robust subgroup discovery and proposes a novel model class and a greedy heuristic algorithm SSD++ to address this problem.
DATA MINING AND KNOWLEDGE DISCOVERY
(2022)
Article
Multidisciplinary Sciences
David Amar, Nasa Sinnott-Armstrong, Euan A. Ashley, Manuel A. Rivas
Summary: cGAUGE is a method for causal inference via Mendelian randomization that can reduce false discovery rate and identify new causal relationships, such as multiple risk factors for cardiovascular disease.
NATURE COMMUNICATIONS
(2021)
Article
Health Care Sciences & Services
Siyun Yang, Ruiwen Zhou, Fan Li, Laine E. Thomas
Summary: This study investigates the propensity score weighting method for causal subgroup survival analysis. Two causal estimands are introduced, and corresponding propensity score weighting estimators are provided. The logistic model with subgroup-covariate interactions selected by least absolute shrinkage and selection operator consistently outperforms other propensity score models. Additionally, overlap weighting generally outperforms inverse probability weighting in terms of balance, bias, and variance.
STATISTICAL METHODS IN MEDICAL RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Konstantina Biza, Ioannis Tsamardinos, Sofia Triantafillou
Summary: Causal discovery algorithms require a set of hyperparameters and selecting the optimal combination is a challenge for practitioners. This study proposes an out-of-sample causal tuning method that treats causal models as predictive models and uses out-of-sample protocols. The method can handle general settings and is evaluated against other tuning approaches, showing its effectiveness in causal discovery.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
W. Qi, A. Abu-Hanna, T. E. M. van Esch, D. de Beurs, Y. Liu, L. E. Flinterman, M. C. Schut
Summary: The objective of the study is to identify patient subgroups with markedly deviating responses to treatment through a modelling approach in observational studies, using pre-treatment variables and Synthetic Random Forest models. The approach was applied to a large primary care dataset and successfully identified four subgroups of positive and negative responders, showing promising predictive value for understanding individual treatment effects.
ARTIFICIAL INTELLIGENCE IN MEDICINE
(2021)
Article
Computer Science, Artificial Intelligence
Yunxia Wang, Fuyuan Cao, Kui Yu, Jiye Liang
Summary: In this article, we propose a new algorithm that utilizes the interventional properties of a causal model to discover the direct causes and direct effects of a target variable from multiple datasets with different manipulations. The algorithm is more suited to real-world cases and tackles a specific challenge. Experimental results validate its effectiveness.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Ana Rita Nogueira, Andrea Pugnana, Salvatore Ruggieri, Dino Pedreschi, Joao Gama
Summary: This article explores the complexity of causality and its significance in the field of artificial intelligence. Causality research aims at obtaining causal knowledge from observational data and estimating the impact of variable changes on outcomes. The article also provides a practical toolkit for researchers and practitioners, including software, datasets, and examples.
WILEY INTERDISCIPLINARY REVIEWS-DATA MINING AND KNOWLEDGE DISCOVERY
(2022)
Article
Statistics & Probability
Chunlin Li, Xiaotong Shen, Wei Pan
Summary: This article introduces a causal discovery method, named DeFuSE, to learn nonlinear relationships in a directed acyclic graph with correlated Gaussian errors due to confounding. DeFuSE includes a deconfounding adjustment to remove the confounding effects and a sequential procedure to estimate the causal order of variables. The proposed approach demonstrates utility and effectiveness in gene regulatory network analysis.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2023)
Article
Computer Science, Interdisciplinary Applications
M. Z. Naser, Arash Teymori Gharah Tapeh
Summary: This paper provides a civil and structural engineering perspective on causal discovery and causal inference. It contrasts common machine learning approaches and outlines key principles and algorithms used in causal discovery and inference. The paper also presents examples and case studies of how causal concepts can be adopted in the domain.
COMPUTERS AND CONCRETE
(2023)
Article
Computer Science, Artificial Intelligence
Gherardo Varando, Salvador Catsis, Emiliano Diaz, Gustau Camps-Valls
Summary: Bivariate causal discovery is the task of inferring the causal relationship between two random variables from observational data. This paper proposes an ensemble algorithm that combines classical and data-driven methods, achieving superior performance on various synthetic and real-world problems.
APPLIED SOFT COMPUTING
(2024)
Article
Engineering, Electrical & Electronic
Kurt Butler, Guanchao Feng, Petar M. Djuric
Summary: Convergent cross mapping is a causal discovery technique for signals that relies on certain assumptions about the underlying systems. This study provides an introduction to the theory of causality, Takens' theorem, and cross maps, and proposes conditions to assess the suitability of a signal for cross mapping. The authors also propose analyses using Gaussian processes to test these conditions in data and demonstrate the detection of potential erroneous results using examples from the literature. They also discuss important considerations when applying convergent cross mapping.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Steffen Kempe, Jochen Hipp, Carsten Lanquillon, Rudolf Kruse
INTERNATIONAL JOURNAL OF UNCERTAINTY FUZZINESS AND KNOWLEDGE-BASED SYSTEMS
(2008)
Article
Automation & Control Systems
Dirk Jacobsen, Carsten Lanquillon, Carsten Wittenberg, Peter Ott