780 Hits in 4.3 sec

Deep Structural Causal Models for Tractable Counterfactual Inference [article]

Nick Pawlowski, Daniel C. Castro, Ben Glocker
2020 arXiv   pre-print
We formulate a general framework for building structural causal models (SCMs) with deep learning components.  ...  The proposed approach employs normalising flows and variational inference to enable tractable inference of exogenous noise variables - a crucial step for counterfactual inference that is missing from existing  ...  Acknowledgements We thank Athanasios Vlontzos for helpful comments on a draft of this paper and the anonymous reviewers for numerous constructive suggestions.  ... 
arXiv:2006.06485v2 fatcat:tysaxdgjd5avvm5l7ovdkimhce

Harmonization with Flow-based Causal Inference [article]

Rongguang Wang, Pratik Chaudhari, Christos Davatzikos
2021 arXiv   pre-print
This paper leverages a recently proposed normalizing-flow-based method to perform counterfactual inference upon a structural causal model (SCM), in order to achieve harmonization of such data.  ...  We infer the posterior of exogenous variables, intervene on observations, and draw samples from the resultant SCM to obtain counterfactuals.  ...  Castro for suggestions. This work was supported by the National Institute on Aging (grant numbers RF1AG054409 and U01AG068057) and the National Institute of Mental Health (grant number R01MH112070).  ... 
arXiv:2106.06845v2 fatcat:hildildhc5go3fojq3rjtcjfuy

On the Tractability of Neural Causal Inference [article]

Matej Zečević and Devendra Singh Dhami and Kristian Kersting
2021 arXiv   pre-print
On another note, research around neural causal models (NCM) recently gained traction, demanding a tighter integration of causality for machine learning.  ...  We prove that SPN-based causal inference is generally tractable, opposed to standard MLP-based NCM.  ...  Now, in the following, we will move onto more general causal models and theoretically investigate tractability of causal inference for these more complex models.  ... 
arXiv:2110.12052v1 fatcat:khgxg2olebhdtnrpladcxyc7xq

Interventions and Counterfactuals in Tractable Probabilistic Models: Limitations of Contemporary Transformations [article]

Ioannis Papantonis, Vaishak Belle
2020 arXiv   pre-print
In this paper, we ask the following technical question: can we use the distributions represented or learned by these models to perform causal queries, such as reasoning about interventions and counterfactuals  ...  For PSDDs the situation is only slightly better. We first provide an algorithm for constructing a causal graph from a PSDD, which introduces augmented variables.  ...  Current structure learning algorithms for tractable models also do not attempt to capture the underlying causal process.  ... 
arXiv:2001.10905v1 fatcat:k5eeedjcffc7pgf65v7jptndzq

A Structural Causal Model for MR Images of Multiple Sclerosis [article]

Jacob C. Reinhold, Aaron Carass, Jerry L. Prince
2021 arXiv   pre-print
These types of questions are causal in nature and require the tools of causal inference to be answered, e.g., with a structural causal model (SCM).  ...  These images can be used for modeling disease progression or used for image processing tasks where controlling for confounders is necessary.  ...  When a DAG has such causal interpretations it is called a structural causal model (SCM), and it represents a generative model of the data on which we can emulate interventions and generate counterfactuals  ... 
arXiv:2103.03158v3 fatcat:rh5l53i26banxggetw6du77ybm

Causal Inference with Deep Causal Graphs [article]

Álvaro Parafita, Jordi Vitrià
2020 arXiv   pre-print
Parametric causal modelling techniques rarely provide functionality for counterfactual estimation, often at the expense of modelling complexity.  ...  We propose Deep Causal Graphs, an abstract specification of the required functionality for a neural network to model causal distributions, and provide a model that satisfies this contract: Normalizing  ...  Section 4 defines the Deep Causal Graph, along with the algorithms to perform causal inference.  ... 
arXiv:2006.08380v1 fatcat:mx4tsyzs4vauzpr6nm4v5i2dy4

Interventional Sum-Product Networks: Causal Inference with Tractable Probabilistic Models [article]

Matej Zečević, Devendra Singh Dhami, Athresh Karanam, Sriraam Natarajan, Kristian Kersting
2021 arXiv   pre-print
While probabilistic models are an important tool for studying causality, doing so suffers from the intractability of inference.  ...  The resulting interventional SPNs are motivated and illustrated by a structural causal model themed around personal health.  ...  ACKNOWLEDGEMENTS The authors thank the anonymous reviewers for their valuable feedback.  ... 
arXiv:2102.10440v5 fatcat:7instaliz5amblf42rdnxdnkay

Deep Causal Graphs for Causal Inference, Black-Box Explainability and Fairness [chapter]

Álvaro Parafita, Jordi Vitrià
2021 Frontiers in Artificial Intelligence and Applications  
We present an alternative framework called Deep Causal Graphs: with a single model, it answers any identifiable causal query without compromising on performance, thanks to the use of Normalizing Causal  ...  Causal Estimation is usually tackled as a two-step process: identification, to transform a causal query into a statistical estimand, and modelling, to compute this estimand by using data.  ...  Vitrià / Deep Causal Graphs for Causal Inference Note that, although f k is deterministic, the effect of E k makes it stochastic w.r.t.  ... 
doi:10.3233/faia210162 fatcat:zauoov6t7vhvxadkl6t7p7lpji

Deep Kalman Filters [article]

Rahul G. Krishnan, Uri Shalit, David Sontag
2015 arXiv   pre-print
Of particular interest is the use of temporal generative models for counterfactual inference.  ...  We investigate the efficacy of such models for counterfactual inference, and to that end we introduce the "Healing MNIST" dataset where long-term structure, noise and actions are applied to sequences of  ...  Acknowledgements The Tesla K40s used for this research were donated by the NVIDIA Corporation.  ... 
arXiv:1511.05121v2 fatcat:pkf3uja5znfvba5lipmwpzcbze

Fairness Through Causal Awareness: Learning Latent-Variable Models for Biased Data [article]

David Madras, Elliot Creager, Toniann Pitassi, Richard Zemel
2018 arXiv   pre-print
Building on prior work in deep learning and generative modeling, we describe how to learn the parameters of this causal model from observational data alone, even in the presence of unobserved confounders  ...  We propose a causal model in which the sensitive attribute confounds both the treatment and the outcome.  ...  BACKGROUND 2.1 Causal Inference We employ Structural Causal Models (SCMs), which provide a general theory for modeling causal relationships between variables [38] .  ... 
arXiv:1809.02519v3 fatcat:4gkvwe4e4ngz7kfe6z257fuvnm

External representations and scientific understanding

Jaakko Kuorikoski, Petri Ylikoski
2014 Synthese  
Finally, the paper shows how the contrastive counterfactual theory of explanation can provide tools for assessing the explanatory power of models.  ...  This paper provides an inferentialist account of model--based understanding by combining a counterfactual account of explanation and an inferentialist account of representation with a view of modeling  ...  We thank the audiences of these events, as well as the reviewers, for their valuable comments.  ... 
doi:10.1007/s11229-014-0591-2 fatcat:5kjvq5htszaptfgfx5hio2txuy

FACEing reality: productive tensions between our epidemiological questions, methods and mission

Nancy Krieger, George Davey Smith
2017 International Journal of Epidemiology  
to the memory of Ruth Hubbard (1924-2016), her mentor since college, who taught her to think critically about science, to do science critically, to challenge whenever biology is invoked as an excuse for  ...  All of the examples share a common point: the need for deep substantive knowledge about the theorized causal processes.  ...  causal inference" [16] .  ... 
doi:10.1093/ije/dyw330 pmid:28130315 fatcat:ebyqlx4jurathimni23fl3pwjy

Counterfactual Adversarial Learning with Representation Interpolation [article]

Wei Wang, Boxin Wang, Ning Shi, Jinfeng Li, Bingyu Zhu, Xiangyu Liu, Rong Zhang
2021 arXiv   pre-print
Deep learning models exhibit a preference for statistical fitting over logical reasoning.  ...  (CRM) on each original-counterfactual pair to adjust sample-wise loss weight dynamically, which encourages the model to explore the true causal effect.  ...  Acknowledgements We appreciate Hui Xue, Shi Chen and Jizhou Kang for sharing their pearls of wisdom.  ... 
arXiv:2109.04746v1 fatcat:56b6uyiytrdjrfrdyaujspgmgq

Causal Reasoning from Meta-reinforcement Learning [article]

Ishita Dasgupta, Jane Wang, Silvia Chiappa, Jovana Mitrovic, Pedro Ortega, David Raposo, Edward Hughes, Peter Battaglia, Matthew Botvinick, Zeb Kurth-Nelson
2019 arXiv   pre-print
The agent can select informative interventions, draw causal inferences from observational data, and make counterfactual predictions.  ...  We train a recurrent network with model-free reinforcement learning to solve a range of problems that each contain causal structure.  ...  Acknowledgements The authors would like to thank the following people for helpful discussions and comments: Neil Rabinowitz, Neil Bramley, Tobias Gerstenberg, Andrea Tacchetti, Victor Bapst, Samuel Gershman  ... 
arXiv:1901.08162v1 fatcat:ie4ifxdojncdrn3axnh74c2ksi

Deep Learning of Potential Outcomes [article]

Bernard Koch, Tim Sainburg, Pablo Geraldo, Song Jiang, Yizhou Sun, Jacob Gates Foster
2021 arXiv   pre-print
for implementing, training, and selecting among deep estimators in Tensorflow 2 available at  ...  This review systematizes the emerging literature for causal inference using deep neural networks under the potential outcomes framework.  ...  Why use deep learning for causal inference?  ... 
arXiv:2110.04442v1 fatcat:kj3dfigqpjeo3j76ykmlephgdq
« Previous Showing results 1 — 15 out of 780 results