Causal Invariance and Counterfactual Learning Driven Cooperative Game for Multi-Label Classification
This addresses challenges in multi-label classification for applications requiring robust and interpretable predictions, especially for rare labels, though it appears incremental by combining existing causal and game theory concepts.
The paper tackled the problem of label imbalance, spurious correlations, and distribution shifts in multi-label classification, particularly for rare label prediction, by introducing the Causal Cooperative Game (CCG) framework, which substantially outperformed strong baselines in rare label prediction and overall robustness.
Multi-label classification (MLC) remains vulnerable to label imbalance, spurious correlations, and distribution shifts, challenges that are particularly detrimental to rare label prediction. To address these limitations, we introduce the Causal Cooperative Game (CCG) framework, which conceptualizes MLC as a cooperative multi-player interaction. CCG unifies explicit causal discovery via Neural Structural Equation Models with a counterfactual curiosity reward to drive robust feature learning. Furthermore, it incorporates a causal invariance loss to ensure generalization across diverse environments, complemented by a specialized enhancement strategy for rare labels. Extensive benchmarking demonstrates that CCG substantially outperforms strong baselines in both rare label prediction and overall robustness. Through rigorous ablation studies and qualitative analysis, we validate the efficacy and interpretability of our components, underscoring the potential of synergizing causal inference with cooperative game theory for advancing multi-label learning.