ECQ$$^{\text {x}}$$: Explainability-Driven Quantization for Low-Bit and Sparse DNNs

Abstract

The remarkable success of deep neural networks (DNNs) in various applications is accompanied by a significant increase in network parameters and arithmetic operations. Such increases in memory and computational demands make deep learning prohibitive for resource-constrained hardware platforms such as mobile devices. Recent efforts aim to reduce these overheads, while preserving model performance as much as possible, and include parameter reduction techniques, parameter quantization, and lossless compression techniques.In this chapter, we develop and describe a novel quantization paradigm for DNNs: Our method leverages concepts of explainable AI (XAI) and concepts of information theory: Instead of assigning weight values based on their distances to the quantization clusters, the assignment function additionally considers weight relevances obtained from Layer-wise Relevance Propagation (LRP) and the information content of the clusters (entropy optimization). The ultimate goal is to preserve the most relevant weights in quantization clusters of highest information content.Experimental results show that this novel Entropy-Constrained and XAI-adjusted Quantization (ECQ$$^{\text {x}}$$ x ) method generates ultra low-precision (2–5 bit) and simultaneously sparse neural networks while maintaining or even improving model performance. Due to reduced parameter precision and high number of zero-elements, the rendered networks are highly compressible in terms of file size, up to 103$$\times $$ × compared to the full-precision unquantized DNN model. Our approach was evaluated on different types of models and datasets (including Google Speech Commands, CIFAR-10 and Pascal VOC) and compared with previous work.

From attribution maps to human-understandable explanations through Concept Relevance Propagation

“…We find, however, that (maximizing) the activation of a latent encoding by a given data point does not always correspond to its utility to the model in an inference context (see, for example, ref. 44 or Supplementary Fig. 7), putting the faithfulness of activation-based example selection for latent concept representation into question.…”

Explainable Artificial Intelligence Applications in Cyber Security: State-of-the-Art in Research

“…On the other hand, the AI algorithms that now often perform best (for example, Deep Learning) are the least explainable, causing a demand for explainable models that can achieve high performance. Some researchers have exploited this area, including authors of [269] significantly reduce the trade-off between efficiency and performance by introducing XAI for DNN into existing quantization techniques. And authors of [270] demonstrated that the wavelet modifications provided could lead to significantly smaller, simplified, more computationally efficient, and more naturally interpretable models, while simultaneously keeping performance.…”