Filters








56 Hits in 7.5 sec

Minimal I-MAP MCMC for Scalable Structure Discovery in Causal DAG Models [article]

Raj Agrawal and Tamara Broderick and Caroline Uhler
2018 arXiv   pre-print
Learning a Bayesian network (BN) from data can be useful for decision-making or discovering causal relationships.  ...  The current state-of-the-art methods such as order MCMC are faster than previous methods but prevent the use of many natural structural priors and still have running time exponential in the maximum indegree  ...  Caroline Uhler was supported in part by NSF (DMS-1651995), ONR (N00014-17-1-2147), and a Sloan Fellowship.  ... 
arXiv:1803.05554v3 fatcat:chf4qviitzenfpm6x66jav2mrq

Variational Causal Networks: Approximate Bayesian Inference over Causal Structures [article]

Yashas Annadani, Jonas Rothfuss, Alexandre Lacoste, Nino Scherrer, Anirudh Goyal, Yoshua Bengio, Stefan Bauer
2021 arXiv   pre-print
Aiming to overcome this issue, we propose a form of variational inference over the graphs of Structural Causal Models (SCMs).  ...  While Bayesian causal inference allows to do so, the posterior over DAGs becomes intractable even for a small number of variables.  ...  A Structural Causal Model (SCM) [29] over X is defined as set of structural assignments X i := f i (x π G (i) , i ) , i = 1, ..., d (1) corresponding to a Directed Acyclic Graph (DAG) G with vertices  ... 
arXiv:2106.07635v1 fatcat:rjiotac6nvhlla6tnv2caoncia

Efficient Sampling and Structure Learning of Bayesian Networks [article]

Jack Kuipers, Polina Suter, Giusi Moffa
2021 arXiv   pre-print
Bayesian networks are probabilistic graphical models widely employed to understand dependencies in high dimensional data, and even to facilitate causal discovery.  ...  Learning the underlying network structure, which is encoded as a directed acyclic graph (DAG) is highly challenging mainly due to the vast number of possible networks in combination with the acyclicity  ...  Acknowledgements The authors would like to thank Markus Kalisch and Niko Beerenwinkel for useful comments and discussions.  ... 
arXiv:1803.07859v4 fatcat:yrqmoqyzenccpik66eab657tou

D'ya Like DAGs? A Survey on Structure Learning and Causal Discovery

Matthew J. Vowels, Necati Cihan Camgoz, Richard Bowden
2022 ACM Computing Surveys  
Causal reasoning is a crucial part of science and human intelligence. In order to discover causal relationships from data, we need structure discovery methods.  ...  We provide a review of background theory and a survey of methods for structure discovery.  ...  Firstly, the model x i = f i (pa i , u i ), i = 1..., n, is a Structural equation/Causal Model (SEM/SCM) which indicates assignment of the value x i in the space of X to a function of its structural parents  ... 
doi:10.1145/3527154 fatcat:sroohzvx5reajkia5ythaiyyjm

Low-variance Black-box Gradient Estimates for the Plackett-Luce Distribution [article]

Artyom Gadetsky, Kirill Struminsky, Christopher Robinson, Novi Quadrianto, Dmitry Vetrov
2019 arXiv   pre-print
To illustrate the approach, we consider a variety of causal structure learning tasks for continuous and discrete data.  ...  In this work, we consider models with latent permutations and propose control variates for the Plackett-Luce distribution.  ...  The authors thank NRU HSE for providing computational resources, NVIDIA for GPU donations, and Amazon for AWS Cloud Credits.  ... 
arXiv:1911.10036v1 fatcat:f7ghfmu2afcq5g7ieb7f5weho4

Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches: Beyond edge weights in psychological networks

Gabor Hullam, Gabriella Juhasz, Bill Deakin, Peter Antal
2017 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB)  
Monte Carlo methods for full Bayesian model averaging over model structures require considerable computational resources, whereas bootstrapped graphical lasso and its approximations offer scalable alternatives  ...  Uncertainty over model structures poses a challenge for many approaches exploring effect strength parameters at system-level.  ...  The sponsors funded the work, but had no further role in the design of the study, in data collection or analysis, in the decision to publish, or in the preparation, review, or approval of the manuscript  ... 
doi:10.1109/cibcb.2017.8058566 dblp:conf/cibcb/HullamJDA17 fatcat:a67mkpbvwbeyzhot3iwxj6tjwi

A Bayesian View of Challenges in Feature Selection: Feature Aggregation, Multiple Targets, Redundancy and Interaction

Peter Antal, András Millinghoffer, Gábor Hullám, Csaba Szalai, András Falus
2008 Journal of machine learning research  
Now we formulate refined levels in BMLA by introducing the concepts of k-MBSs and k-MBGs, which are intermediate, scalable model properties expressing relevance.  ...  We introduce and investigate a score for feature redundancy and interaction based on the decomposability of the structure posterior.  ...  In ( 16 ), Madigan et al. proposed a Markov Chain Monte Carlo (MCMC) scheme to approximate such Bayesian inference. The MCMC method over the DAG space was improved by Castelo et al. (12) .  ... 
dblp:journals/jmlr/AntalMHSF08 fatcat:dhy62s5a3vbmpewohgayblcz74

pg-Causality: Identifying Spatiotemporal Causal Pathways for Air Pollutants with Urban Big Data [article]

Julie Yixuan Zhu, Chao Zhang, Huichu Zhang, Shi Zhi, Victor O.K. Li, Jiawei Han, Yu Zheng
2018 arXiv   pre-print
Then, Bayesian learning carefully encodes the local and ST causal relations with a Gaussian Bayesian network (GBN)-based graphical model, which also integrates environmental influences to minimize biases  ...  In this paper, we use urban big data (air quality data and meteorological data) to identify the spatiotemporal (ST) causal pathways for air pollutants.  ...  Pearl's causality model [5] encodes the causal relationships in a directed acyclic graph (DAG) [23] for probabilistic inference.  ... 
arXiv:1610.07045v3 fatcat:vs22xf233rfqbaddofx6muw2ze

A survey of Bayesian Network structure learning [article]

Neville K. Kitson, Anthony C. Constantinou, Zhigao Guo, Yang Liu, Kiattikun Chobtham
2021 arXiv   pre-print
However, determining the graphical structure of a BN remains a major challenge, especially when modelling a problem under causal assumptions.  ...  Approaches for dealing with data noise in real-world datasets and incorporating expert knowledge into the learning process are also covered.  ...  Pearl (1998) showed that this creates a minimal I-MAP DAG for that node ordering.  ... 
arXiv:2109.11415v1 fatcat:tj4ceig7rvbb3k4bqffvdvvwye

pg-Causality: Identifying Spatiotemporal Causal Pathways for Air Pollutants with Urban Big Data

Julie Yixuan Zhu, Chao Zhang, Huichu Zhang, Shi Zhi, Victor O.K. Li, Jiawei Han, Yu Zheng
2017 IEEE Transactions on Big Data  
Then, Bayesian learning carefully encodes the local and ST causal relations with a Gaussian Bayesian Network (GBN)-based graphical model, which also integrates environmental influences to minimize biases  ...  In this paper, we use urban big data (air quality data and meteorological data) to identify the spatiotemporal (ST) causal pathways for air pollutants.  ...  Pearl's causality model [5] encodes the causal relationships in a directed acyclic graph (DAG) [23] for probabilistic inference.  ... 
doi:10.1109/tbdata.2017.2723899 fatcat:kf4hq6yygrdizj6ift5m3lfhz4

A Bayesian Active Learning Experimental Design for Inferring Signaling Networks

Robert O. Ness, Karen Sachs, Parag Mallick, Olga Vitek
2018 Journal of Computational Biology  
To learn the causal patterns of influence between proteins in a regulatory network, the methods require experiments that include targeted interventions.  ...  Our approach takes as input pathway databases such as KEGG and historic datasets in repositories such as Cytobank, expresses them in form of prior probability distributions on the signaling network structures  ...  Scutari for guidance in using the R package bnlearn. This work was supported in part by the NSF CAREER award DBI-1054826, and by the Sy and Laurie Sternberg award to OV.  ... 
doi:10.1089/cmb.2017.0247 pmid:29927613 fatcat:humwqmytfbexze3weng2pejvxi

Causal Queries from Observational Data in Biological Systems via Bayesian Networks: An Empirical Study in Small Networks [article]

Alex White, Matthieu Vignes
2018 arXiv   pre-print
We go one step further and advocate for the consideration of causal representations of the interactions in living systems.We present the causal formalism and bring it out in the context of biological networks  ...  We also discuss its ability to decipher the causal information flow as observed in gene expression.  ...  Lastly, we are very grateful to reviewers of this chapter for their insightful comments.  ... 
arXiv:1805.01608v1 fatcat:gcvsm6pkgvhmhnswo5vb22fkbe

High-throughput Bayesian network learning using heterogeneous multicore computers

Michael D. Linderman, Robert Bruggner, Vivek Athalye, Teresa H. Meng, Narges Bani Asadi, Garry P. Nolan
2010 Proceedings of the 24th ACM International Conference on Supercomputing - ICS '10  
The causal structure of signal transduction networks can be modeled as Bayesian Networks (BNs), and computationally learned from experimental data.  ...  However, learning the structure of Bayesian Networks (BNs) is an NPhard problem that, even with fast heuristics, is too time consuming for large, clinically important networks (20-50 nodes).  ...  and David Dill for their help with precision analysis.  ... 
doi:10.1145/1810085.1810101 pmid:28819655 pmcid:PMC5557010 dblp:conf/ics/LindermanBAMAN10 fatcat:jtrq4z3fdzdtbfh6svqu7yepkq

DAGs with No Curl: An Efficient DAG Structure Learning Approach [article]

Yue Yu, Tian Gao, Naiyu Yin, Qiang Ji
2021 arXiv   pre-print
in this equivalent set of DAGs.  ...  To further improve efficiency, we propose a novel learning framework to model and learn the weighted adjacency matrices in the DAG space directly.  ...  Changhe Yuan for helpful discussions. We thank to anonymous reviewers who have provide helpful comments. Y.  ... 
arXiv:2106.07197v1 fatcat:4bxn3ofwlfe4bj7icfk2wj4uqe

A hierarchical Bayesian network approach for linkage disequilibrium modeling and data-dimensionality reduction prior to genome-wide association studies

Raphaël Mourad, Christine Sinoquet, Philippe Leray
2011 BMC Bioinformatics  
In order to tackle the learning of both forest structure and probability distributions, a generic algorithm has been proposed.  ...  Conclusions: The forest of hierarchical latent class models offers several advantages for genome-wide association studies: accurate modeling of linkage disequilibrium, flexible data dimensionality reduction  ...  Acknowledgements The authors are grateful to two anonymous referees for constructive comments and help in improving their manuscript.  ... 
doi:10.1186/1471-2105-12-16 pmid:21226914 pmcid:PMC3033325 fatcat:dbgrxmj46zcn5cr52e2x47hrsa
« Previous Showing results 1 — 15 out of 56 results