Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Causally Fair Node Classification on Non-IID Graph Data

Yucong Dai, Lu Zhang, Yaowei Hu, Susan Gauch, Yongkai Wu

TL;DR

This work tackles fairness in non-IID graph data by extending causal inference to networks via the Network Structural Causal Model (NSCM) and two key assumptions—Decomposability and Graph Independence—which enable $do$-calculus-based interventions on graphs. It introduces MPVA, a hybrid of a Message Passing Neural Network (MPNN) and a conditional Variational Autoencoder (cVAE), to learn interventional distributions $P(x|do( ext{s}))$ under graph interference and to support a causally fair node classifier through a regularization term that aligns outcomes across favorable and unfavorable groups. Empirical results on semi-synthetic and real graphs show MPVA better approximates non-IID interventional effects (gCF) and mitigates bias with competitive accuracy compared to baselines like FairGNN and GEAR. The findings underscore the potential of causality-based fairness in complex graph applications and point to future work on relaxing assumptions to broaden applicability.

Abstract

Fair machine learning seeks to identify and mitigate biases in predictions against unfavorable populations characterized by demographic attributes, such as race and gender. Recently, a few works have extended fairness to graph data, such as social networks, but most of them neglect the causal relationships among data instances. This paper addresses the prevalent challenge in fairness-aware ML algorithms, which typically assume Independent and Identically Distributed (IID) data. We tackle the overlooked domain of non-IID, graph-based settings where data instances are interconnected, influencing the outcomes of fairness interventions. We base our research on the Network Structural Causal Model (NSCM) framework and posit two main assumptions: Decomposability and Graph Independence, which enable the computation of interventional distributions in non-IID settings using the $do$-calculus. Based on that, we develop the Message Passing Variational Autoencoder for Causal Inference (MPVA) to compute interventional distributions and facilitate causally fair node classification through estimated interventional distributions. Empirical evaluations on semi-synthetic and real-world datasets demonstrate that MPVA outperforms conventional methods by effectively approximating interventional distributions and mitigating bias. The implications of our findings underscore the potential of causality-based fairness in complex ML applications, setting the stage for further research into relaxing the initial assumptions to enhance model fairness.

Causally Fair Node Classification on Non-IID Graph Data

TL;DR

This work tackles fairness in non-IID graph data by extending causal inference to networks via the Network Structural Causal Model (NSCM) and two key assumptions—Decomposability and Graph Independence—which enable -calculus-based interventions on graphs. It introduces MPVA, a hybrid of a Message Passing Neural Network (MPNN) and a conditional Variational Autoencoder (cVAE), to learn interventional distributions under graph interference and to support a causally fair node classifier through a regularization term that aligns outcomes across favorable and unfavorable groups. Empirical results on semi-synthetic and real graphs show MPVA better approximates non-IID interventional effects (gCF) and mitigates bias with competitive accuracy compared to baselines like FairGNN and GEAR. The findings underscore the potential of causality-based fairness in complex graph applications and point to future work on relaxing assumptions to broaden applicability.

Abstract

Fair machine learning seeks to identify and mitigate biases in predictions against unfavorable populations characterized by demographic attributes, such as race and gender. Recently, a few works have extended fairness to graph data, such as social networks, but most of them neglect the causal relationships among data instances. This paper addresses the prevalent challenge in fairness-aware ML algorithms, which typically assume Independent and Identically Distributed (IID) data. We tackle the overlooked domain of non-IID, graph-based settings where data instances are interconnected, influencing the outcomes of fairness interventions. We base our research on the Network Structural Causal Model (NSCM) framework and posit two main assumptions: Decomposability and Graph Independence, which enable the computation of interventional distributions in non-IID settings using the -calculus. Based on that, we develop the Message Passing Variational Autoencoder for Causal Inference (MPVA) to compute interventional distributions and facilitate causally fair node classification through estimated interventional distributions. Empirical evaluations on semi-synthetic and real-world datasets demonstrate that MPVA outperforms conventional methods by effectively approximating interventional distributions and mitigating bias. The implications of our findings underscore the potential of causality-based fairness in complex ML applications, setting the stage for further research into relaxing the initial assumptions to enhance model fairness.
Paper Structure (28 sections, 7 theorems, 16 equations, 7 figures, 3 tables)

This paper contains 28 sections, 7 theorems, 16 equations, 7 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminary
  3. Problem Formulation
  4. Network Structural Causal Model (NSCM)
  5. Fair Node Classification
  6. Proposed Research
  7. Causal Inference on Network Data
  8. Message Passing Variational Autoencoder for Causal Inference (MPVA)
  9. Causally Fair Node Classification
  10. Experiment
  11. Datasets
  12. Experiment Settings
  13. Fairness Metrics:
  14. Mitigation Baselines:
  15. Results on Semi-synthetic Data
  16. ...and 13 more sections

Key Result

Lemma 1

For two different nodes $i,j$, $c^{(i)}=c^{(j)}$ if and only if nodes $i$ and $j$ have identical computation trees in the WL graph isomorphism test.

Figures (7)

  • Figure 1: Graphs and diagrams used in this paper.
  • Figure 2: Networked causal diagram for node classification.
  • Figure 3: The causal diagram that is equivalent to the networked causal diagram in Fig. \ref{['fig:fnc']}.
  • Figure 4: The MPVA framework. MPNN learns the hidden representation of aggregated causal effects from neighbors through reconstruction. cVAE learns the conditional distribution for computing the interventional distribution.
  • Figure 5: Comparison of measured bias and true gCF bias on D2 with various mitigation methods.
  • ...and 2 more figures

Theorems & Definitions (10)

  • Definition 1: Network Structural Causal Model (NSCM)
  • Lemma 1: DBLP:journals/corr/abs-2204-07697
  • Proposition 2
  • proof
  • Theorem 3
  • Lemma 4
  • Lemma 5
  • Lemma 6
  • Theorem 3
  • proof