Article
Quantum Science & Technology
Cristhian Roman-Vicharra, James J. Cai
Summary: In this study, a quantum circuit model was proposed to infer gene regulatory networks (GRNs) from single-cell transcriptomic data. The model utilized qubit entanglement to simulate interactions between genes, showing competitive performance and potential for further exploration. The application of the quantum GRN modeling approach to human lymphoblastoid cells successfully predicted regulatory interactions between genes and estimated the strength of these interactions. This work highlights the potential of quantum computing in biology for a better understanding of single-cell GRNs.
NPJ QUANTUM INFORMATION
(2023)
Article
Psychology, Mathematical
Don van den Bergh, Merlise A. Clyde, Akash R. Komarlu Narendra Gupta, Tim de Jong, Quentin F. Gronau, Maarten Marsman, Alexander Ly, Eric-Jan Wagenmakers
Summary: Linear regression analysis commonly involves two stages: defining the best model and using regression coefficients for prediction and evaluation; traditional inference methods often ignore model uncertainty, leading to overconfident parameter estimates. Model averaging is a technique that overcomes these drawbacks by weighting the contribution of each model for inference.
BEHAVIOR RESEARCH METHODS
(2021)
Article
Computer Science, Interdisciplinary Applications
David Reynolds, Luis Carvalho
Summary: A novel approach to statistical inference for multivariate binary transaction data is proposed in this study. The hierarchical model and MCMC sampling procedure provided a sparser representation of item associations compared to frequent itemset mining, without sacrificing predictive accuracy. By allowing inference on a broad set of parameters, the model offers a deeper level of insight into transaction data.
COMPUTATIONAL STATISTICS & DATA ANALYSIS
(2021)
Article
Physics, Multidisciplinary
Shouta Sugahara, Itsuki Aomi, Maomi Ueno
Summary: This study aims to improve classification accuracy using a modified subbagging method and demonstrates that it outperforms previous methods in the field of Bayesian network classification.
Article
Biochemical Research Methods
Polina Suter, Jack Kuipers, Niko Beerenwinkel
Summary: This study presents a strategy for learning gene regulatory networks (GRNs) using Dynamic Bayesian networks (DBNs) from gene expression data. The proposed approach is scalable, has high predictive accuracy, and prevents overfitting. The application of DBNs to two time series transcriptomic datasets demonstrates improved classification accuracy and the identification of differences in gene networks between cancer and normal tissues.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Biochemical Research Methods
Alexander P. Wu, Jian Peng, Bonnie Berger, Hyunghoon Cho
Summary: ShareNet is a Bayesian framework that enhances the accuracy of cell type-specific gene regulatory networks by sharing information across related cell types. The method demonstrates significantly improved accuracy on three benchmark scRNA-seq datasets and reveals key changes in gene associations that support the complex rewiring of regulatory networks.
Article
Biochemical Research Methods
Qinhuan Luo, Yongzhen Yu, Xun Lan
Summary: SIGNET is a deep learning-based framework that captures complex regulatory relationships between genes. Evaluations on various real and simulated scRNA-seq datasets showed that SIGNET is more sensitive to regulatory interactions in different types of cells, especially rare cells. Therefore, SIGNET is a useful tool for downstream analyses such as cell clustering and gene regulatory network inference.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Automation & Control Systems
Jeremias Knoblauch, Jack Jewson, Theodoros Damoulas
Summary: The paper advocates for an optimization-centric view of Bayesian inference, introducing the Rule of Three (ROT) as a generalized method for Bayesian posteriors. It also explores the applications of Generalized Variational Inference (GVI) posteriors and their potential to improve robustness and posterior marginals in Bayesian Neural Networks and Deep Gaussian Processes.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Multidisciplinary Sciences
Frantisek Bartos, Maximilian Maier, David R. Shanks, T. D. Stanley, Martina Sladekova, Eric-Jan Wagenmakers
Summary: Adjusting for publication bias is important in meta-analytic inferences. Most existing methods do not work well under various research conditions. Sladekova et al. proposed an approach to select the most appropriate methods for specific conditions and found that publication bias only slightly overestimates effect sizes in psychology on average. However, their approach faces a 'Catch-22' problem, which can be mitigated by using robust Bayesian meta-analysis.
ROYAL SOCIETY OPEN SCIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Martin Magris, Alexandros Iosifidis
Summary: The last decade has seen a growing interest in Bayesian learning, but its technicality and complexity in practical implementations have limited its widespread adoption. This survey introduces the principles and algorithms of Bayesian Learning for Neural Networks from a practical perspective, discussing standard and recent approaches for Bayesian inference. It also explores the use of manifold optimization as a state-of-the-art approach and provides pseudo-codes for implementation.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Review
Health Care Sciences & Services
Sarah Friedrich, Andreas Groll, Katja Ickstadt, Thomas Kneib, Markus Pauly, Joerg Rahnenfuhrer, Tim Friede
Summary: This article reviews regularization approaches in data science for overcoming overfitting and improving prediction, and discusses their limited application in medical research. The authors suggest increased use of regularization approaches in medicine, despite the added complexity they bring to analyses. Proper investments in computing facilities and educational resources can help overcome these challenges.
STATISTICAL METHODS IN MEDICAL RESEARCH
(2023)
Article
Engineering, Civil
Zijing Xie, Yunjun Yao, Xiaotong Zhang, Shunlin Liang, Joshua B. Fisher, Jiquan Chen, Kun Jia, Ke Shang, Junming Yang, Ruiyang Yu, Xiaozheng Guo, Lu Liu, Jing Ning, Lilin Zhang
Summary: In this study, a deep neural network merging framework was introduced to improve the estimation of global terrestrial evapotranspiration (ET). The results showed that the deep neural network method had higher accuracy compared to other methods. The global terrestrial ET product generated based on this method can be used to monitor regional and global water resources and environmental changes.
JOURNAL OF HYDROLOGY
(2022)
Article
Biochemical Research Methods
Julia Akesson, Zelmina Lubovac-Pilav, Rasmus Magnusson, Mika Gustafsson
Summary: ComHub is a tool for predicting hubs in GRNs by averaging predictions from a compendium of network inference methods. Benchmarking against DREAM5 challenge data and independent gene expression datasets showed robust performance. ComHub consistently scored among the top performing methods on data from different sources, and can work with both predefined networks and perform stand-alone network inference, making it generally applicable.
BMC BIOINFORMATICS
(2021)
Article
Mathematics
Francisco Javier Diez, Manuel Arias, Jorge Perez-Martin, Manuel Luque
Summary: OpenMarkov is an open-source software tool designed for probabilistic graphical models, primarily in medicine but also used in other fields and education in over 30 countries. This paper explains how OpenMarkov can be used as a pedagogical tool to teach the main concepts of Bayesian networks and influence diagrams, as well as various inference algorithms.
Article
Mathematical & Computational Biology
Jeffrey A. Boatman, David M. Vock, Joseph S. Koopmeiners
Summary: The increasing diversity of data sources offers more possibilities for estimating treatment effects, but borrowing must be done in a principled manner to reduce bias and errors. Estimators based on regression and Bayesian methods show promising performance in handling causal effects and can be applied to different data sources.
STATISTICS IN MEDICINE
(2021)
Article
Statistics & Probability
Yicheng Li, Adrian E. Raftery
Summary: The study introduces a new method to improve the quality of mortality forecasts by accounting for the impact of smoking on life expectancy. By utilizing a Bayesian hierarchical model and age-specific smoking attributable fraction, the forecast of nonsmoking life expectancy at birth is converted into life expectancy forecast. The proposed method demonstrates improvements in forecast accuracy and model calibration compared to other commonly used methods for life expectancy forecasting.
ANNALS OF APPLIED STATISTICS
(2021)
Article
Statistics & Probability
Bei Jiang, Adrian E. Raftery, Russell J. Steele, Naisyin Wang
Summary: There is a growing expectation for government-funded studies to make collected data openly available to ensure reproducibility, but concerns about data privacy also arise. One strategy to protect privacy is to release synthetic datasets with masked sensitivity values, but this can lead to information loss or invalid inferences. A new masking framework with a data-augmentation component has been proposed to alleviate these issues.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2021)
Article
Statistics & Probability
Hannah M. Director, Adrian E. Raftery, Cecilia M. Bitz
Summary: This paper introduces a method to probabilistically forecast sea ice using a mixture of ensemble models and observed sea ice patterns. The forecasts produced by this method are better calibrated than other methods at short lead times.
ANNALS OF APPLIED STATISTICS
(2021)
Article
Biology
Peter A. Gao, Hannah M. Director, Cecilia M. Bitz, Adrian E. Raftery
Summary: This study focuses on the impact of warming temperatures on the volume of sea ice in the Arctic Ocean, proposing a statistical spatio-temporal two-stage model for predicting sea ice thickness. By combining a contour model and a Gaussian random field, the model is able to generate probabilistic forecasts up to three months into the future, showing comparable accuracy and improved calibration compared to existing forecasts. The statistical model also demonstrates the ability to generate good forecasts of aggregate quantities such as overall and regional sea ice volume.
JOURNAL OF AGRICULTURAL BIOLOGICAL AND ENVIRONMENTAL STATISTICS
(2022)
Article
Economics
Adrian E. Raftery, Hana Sevcikova
Summary: Population forecasts are important for governments and the private sector for planning purposes. Traditional deterministic methods are being replaced with probabilistic forecasts to assess accuracy and make risk-based decisions. The United Nations has made significant progress in issuing probabilistic population forecasts for all countries using a Bayesian methodology. Additionally, long-term population and economic forecasts are crucial for assessing the social cost of carbon emissions.
INTERNATIONAL JOURNAL OF FORECASTING
(2023)
Article
Multidisciplinary Sciences
Anupreet Porwal, Adrian E. Raftery
Summary: Probability models are widely used in statistical tasks and it is important to choose an appropriate model and consider the uncertainty associated with this choice. This study focuses on variable selection in linear regression models and compares 21 popular methods through simulation studies. The results show that three adaptive Bayesian model averaging (BMA) methods perform the best across all statistical tasks.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2022)
Article
Meteorology & Atmospheric Sciences
Xin Chen, Adrian E. Raftery, David S. Battisti, Peiran R. Liu
Summary: This study developed a probabilistic method for predicting long-term spatial temperature changes, showing that high latitudes, continents, and the Northern Hemisphere warm more quickly. The Arctic may experience a temperature increase of up to 16 degrees Celsius, while North Africa, West Asia, and most of Europe are expected to see a temperature rise of at least 2 degrees Celsius.
Article
Statistics & Probability
Hannah M. Director, Adrian E. Raftery
Summary: This article introduces a method for modeling spatial point sequences as boundaries, such as the sea ice edge, by generating sets of distances in different directions. Metrics are introduced to assess the deviation of boundary shapes from star-shapedness.
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES C-APPLIED STATISTICS
(2022)
Article
Multidisciplinary Sciences
Kevin Rennert, Frank Errickson, Brian C. Prest, Lisa Rennels, Richard G. Newell, William Pizer, Cora Kingdon, Jordan Wingenroth, Roger Cooke, Bryan Parthum, David Smith, Kevin Cromar, Delavane Diaz, Frances C. Moore, Ulrich K. Muller, Richard J. Plevin, Adrian E. Raftery, Hana Sevcikova, Hannah Sheets, James H. Stock, Tammy Tan, Mark Watson, Tony E. Wong, David Anthoff
Summary: This study shows that improved probabilistic socioeconomic projections, climate models, damage functions, and discounting methods can significantly increase the estimates of the social cost of carbon dioxide (SC-CO2). The study's estimates are higher than the current values used in policy evaluation, thereby increasing the expected benefits of greenhouse gas mitigation.
Article
Multidisciplinary Sciences
Nathan G. Welch, Adrian E. Raftery
Summary: We propose a method for forecasting global human migration flows using a Bayesian hierarchical model. Our method produces well-calibrated out-of-sample forecasts and accurately assesses uncertainty. It improves on existing methods and provides accurate projections of migration flows.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2022)
Article
Mathematics, Interdisciplinary Applications
Isobel Claire Gormley, Thomas Brendan Murphy, Adrian E. Raftery
Summary: Clustering is an automatic task to group observations into homogeneous groups, and the number of groups is unknown. Model-based clustering provides a principled and reproducible approach to clustering based on statistical modeling framework. It allows for robust parameter estimation and objective inference on the number of clusters, while considering uncertainty in cluster membership.
ANNUAL REVIEW OF STATISTICS AND ITS APPLICATION
(2023)
Article
Environmental Sciences
Lucas R. Vargas Zeppetello, Adrian E. Raftery, David S. Battisti
Summary: Using probabilistic emission projections, this study demonstrates that the Heat Index driven by anthropogenic CO2 emissions will increase global exposure to dangerous environments in the future. Even if the Paris Agreement goal of limiting global warming to 2 degrees C is achieved, the exposure to dangerous Heat Index levels is projected to rise significantly, particularly in tropical regions. Without more aggressive emissions reductions, it is predicted that by 2100, many people in tropical regions will experience dangerously high Heat Index values on a daily basis, and deadly heat waves will become annual occurrences in mid-latitude regions.
COMMUNICATIONS EARTH & ENVIRONMENT
(2022)
Article
Social Sciences, Mathematical Methods
J. Mulder, A. E. Raftery
Summary: This article introduces a method of evaluating models with order constraints, called order-constrained Bayesian information criterion (BIC), which improves the traditional BIC by using truncated information priors. Experimental results show that the order-constrained BIC based on the local unit information prior has better performance and lower error probabilities when evaluating order-constrained models.
SOCIOLOGICAL METHODS & RESEARCH
(2022)
Article
Economics
Kevin Rennert, Brian C. Prest, William A. Pizer, Richard G. Newell, David Anthoff, Cora Kingdon, Lisa Rennels, Roger Cooke, Adrian E. Raftery, Hana Sevcikova, Frank Errickson
Summary: The social cost of carbon is a crucial metric for climate policy, requiring consideration of uncertainty and transparency of assumptions. Challenges in estimation push the boundaries of analytical techniques, necessitating augmented approaches to assess uncertainty and discounting.
BROOKINGS PAPERS ON ECONOMIC ACTIVITY
(2021)
Article
Computer Science, Interdisciplinary Applications
Adrian E. Raftery, Hana Sevcikova, Bernard W. Silverman
Summary: The vote package in R describes methods for selecting the Condorcet winner and loser, emphasizing the practical application of the STV system in small electorates. It also introduces a new variant of STV and illustrates the package with real examples.