Generalizability through Explainability: Countering Overfitting with Counterfactual Examples

Overfitting is a well-known issue in machine learning that occurs when a model struggles to generalize its predictions to new, unseen data beyond the scope of its training set. Traditional techniques to mitigate overfitting include early stopping, data augmentation, and regularization. In this work, we demonstrate that the degree of overfitting of a trained model is correlated with the ability to generate counterfactual examples. The higher the overfitting, the easier it will be to find a valid counterfactual example for a randomly chosen input data point. Therefore, we introduce CF-Reg, a novel regularization term in the training loss that controls overfitting by ensuring enough margin between each instance and its corresponding counterfactual. Experiments conducted across multiple datasets and models show that our counterfactual regularizer generally outperforms existing regularization techniques.

Effective Non-Random Extreme Learning Machine

The Extreme Learning Machine (ELM) is a growing statistical technique widely applied to regression problems. In essence, ELMs are single-layer neural networks where the hidden layer weights are randomly sampled from a specific distribution, while the output layer weights are learned from the data. Two of the key challenges with this approach are the architecture design, specifically determining the optimal number of neurons in the hidden layer, and the method's sensitivity to the random initialization of hidden layer weights. This paper introduces a new and enhanced learning algorithm for regression tasks, the Effective Non-Random ELM (ENR-ELM), which simplifies the architecture design and eliminates the need for random hidden layer weight selection. The proposed method incorporates concepts from signal processing, such as basis functions and projections, into the ELM framework. We introduce two versions of the ENR-ELM: the approximated ENR-ELM and the incremental ENR-ELM. Experimental results on both synthetic and real datasets demonstrate that our method overcomes the problems of traditional ELM while maintaining comparable predictive performance.