Harnessing the Power of Explanations for Incremental Training: A LIME-Based Approach

TL;DR

This paper introduces a LIME-based feedback mechanism integrated into neural network training to improve incremental learning performance, especially in scenarios with sequential data and limited access to previous data.

Contribution

It proposes a novel weighted loss based on LIME explanations combined with Elastic Weight Consolidation to enhance incremental training of neural networks.

Findings

Achieved 0.5% to 1.5% accuracy improvement in keyword spotting.

Demonstrated effective model generalization with explanation-guided training.

Validated approach on Google Speech Commands dataset.

Abstract

Explainability of neural network prediction is essential to understand feature importance and gain interpretable insight into neural network performance. However, explanations of neural network outcomes are mostly limited to visualization, and there is scarce work that looks to use these explanations as feedback to improve model performance. In this work, model explanations are fed back to the feed-forward training to help the model generalize better. To this extent, a custom weighted loss where the weights are generated by considering the Euclidean distances between true LIME (Local Interpretable Model-Agnostic Explanations) explanations and model-predicted LIME explanations is proposed. Also, in practical training scenarios, developing a solution that can help the model learn sequentially without losing information on previous data distribution is imperative due to the unavailability of all the training data at once. Thus, the framework incorporates the custom weighted loss with Elastic Weight Consolidation (EWC) to maintain performance in sequential testing sets. The proposed custom training procedure results in a consistent enhancement of accuracy ranging from 0.5% to 1.5% throughout all phases of the incremental learning setup compared to traditional loss-based training methods for the keyword spotting task using the Google Speech Commands dataset.