On Estimation of Functional Causal Models

By Kun Zhang, Zhikun Wang, Jiji Zhang, Bernhard Schölkopf. 2015

109 citations recorded by the source corpus.

Subfields: Analytical Chemistry, Artificial Intelligence, Probabilistic & Bayesian Methods, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Analytical Chemistry

1. Clark Glymour
2. Peter Spirtes
3. Biwei Huang
4. Bernhard Schölkopf
5. Eric P. Xing
6. Kun Zhang
7. Jiji Zhang
8. Pradeep Ravikumar

In Artificial Intelligence

1. Bernhard Schölkopf
2. Simon Lacoste-Julien
3. Kun Zhang
4. Clark Glymour
5. Shengyu Zhu
6. Ignavier Ng
7. Zhitang Chen
8. Biwei Huang

In Probabilistic & Bayesian Methods

1. Bernhard Schölkopf
2. Kun Zhang
3. Eric P. Xing
4. Peter Spirtes
5. Simon Lacoste-Julien
6. Clark Glymour
7. Biwei Huang
8. Ignavier Ng

In Signal Processing

1. Bernhard Schölkopf
2. Clark Glymour
3. Kun Zhang
4. Jiji Zhang
5. Peter Spirtes
6. Biwei Huang
7. Ignavier Ng
8. Shengyu Zhu

Citations in the corpus

• Gradient-Based Neural DAG Learning (2019, methodology): “Our first contribution is to extend the framework of Zheng et al. (2018) to deal with nonlinear relationships between variables using neural networks (NN) (Goodfellow et al.”
• Sample-Efficient Reinforcement Learning via Counterfactual-Based Data Augmentation (2020, background): “Furthermore, note that although the function f and the probabilistic distribution P (Ut+1) are not uniquely identifiable from the collected data without strong constraints on the functional class of f and on the data distributions [20], surprisingly, the constructed counterfactual outcome does not d”
• Review of Causal Discovery Methods Based on Graphical Models (2019, background): “In the presence of outcome-dependent selection bias, with FCM-based causal discovery it is possible to identify the correct causal direction and estimate properties of the causal mechanism (Zhang et al., 2016).”
• Distinguishing Cause from Effect Based on Exogeneity (2015, background): “Recently, causal discovery based on structural equation models (SEMs) has proved useful in distinguishing cause from effect [21, 8, 27, 26, 15, 29]; however, the performance of such approaches depends on assumptions on the functional model class and/or on the data-generating functions.”
• Learning Sparse Nonparametric DAGs (2019, methodology): “…additive noise models (Hoyer et al., 2009; Peters et al., 2014; Blöbaum et al., 2018; Mooij et al., 2016), post-nonlinear models (Zhang and Hyvärinen, 2009; Zhang et al., 2016) and general nonlinear SEM (Monti et al., 2019; Goudet et al., 2018; Kalainathan et al., 2018; Sgouritsa et al., 2015).”
• Explainable and Interpretable Models in Computer Vision and Machine Learning (2017, unknown): “In the absence of any such assumption, Zhang et al. (2016) show that, even in the bivariate case, for any function f and noise variable E independent of X such that Y = f ( X , E ) , it is always feasible to construct some ˜ f and ˜ E , with ˜ E independent of Y , such that X = ˜ f ( Y , ˜ E ) .”
• Masked Gradient-Based Causal Structure Learning (2019, methodology): “In general, however, it is impossible to recover G using only observational data from P ( X ), because the underlying DAG is generally not identifiable without further assumption on the SEM [25, 41].”