Disentangled Explanations for Neural Network Predictions on Audio Data

This page depicts the results of the experiments conducted within the scope of my master's thesis which I completed at the Machine Learning Group TU Berlin under the supervision of K. R. Müller and Grégoire Montavon. The abstract and a brief overview of the research process are provided in the following. For a comprehensive overview of the methodologies implemented in this work, please refer to the report. Code and report are linked below.

Relevance Decomposition with DRSA
Fig. 1: This figure shows the virtual layers \((U_k)_k\) trained with DRSA, that map activations \(a_j\) onto relevant subspaces \(\textbf{h}_k\). The black arrows depict the forward pass of some input sample trough the neural network, resulting in evidence for some class (here \(y_4\) for demonstration purposes). Red arrows show the relevance flow of the class evidence of \(y_4\) to the inputs as implemented with LRP. By keeping the relevance flow disentangled after filtering it through the subspaces, distinct explanation components can be visualized at the inputs.

Abstract

As nonlinear Machine Learning (ML) models are increasingly used in various real world applications, their black-box nature hinders in-depth model evaluation, apart from performance measures. In response, the field of Explainable Artificial Intelligence (XAI) has made much progress. It aims to reveal the rationale behind complex ML models, often by assigning relevance scores to model parts and input features, e.g., pixels. However, in some domains such as audio processing, where data—like time or time-frequency representations of amplitudes—is of rather unintuitive nature, the extracted explanations can be hard to grasp for humans. Suitably, a novel sub-field in the realm of XAI has emerged in very recent time that aims to decompose explanations into multiple sub-explanations, representing distinct decision concepts. These approaches offer a promising foundation to gain more valuable insights into models and the data domain, especially when dealing with complex data scenarios. This study targets the extraction of concept-based explanations for a neural network applied to audio classification tasks, by utilizing the newly proposed method, Disentangled Relevant Subspace Analysis (DRSA), in combination with Layerwise Relevance Propagation (LRP).

View Report View Code

Brief Summary of Research Methodology and Process

  1. Conceptualization and training of a Deep Learning (DL) approach to solve audio classification tasks on inputs in time-frequency domain. The tasks are solved with a Convolutional Neural Network. The main showcase is conducted on a Music Genre Recognition (MGR) task on the GTZAN dataset [1].
  2. Explanation of model decisions on a local basis by utilizing the backpropagation-based XAI method LRP [2].
  3. Subsequently, relevant subspaces are optimized in latent space of the network with the newly proposed concept-based XAI method DRSA [3]. By implementing a two-step attribution that allows to condition relevances on single subspaces while propagating those from the outputs to the inputs, concept-based explanations can be visualized in the input domain. This process is schematically depicted in the figure below.
  4. Design of a methodology to transform explanations in time-frequency space into listenable audios to maximize human interpretability of those.
  5. Qualitative and quantitative evaluation of the extracted disentangled explanations to demonstrate the advantage of concept-based explanations and the enhanced explanatory value obtained through the application of DRSA. To provide a solid basis for evaluation, a synthetic dataset is created that fits the purpose of this study.




Explore Results

Music Showcase Synthetic Toy Case



References

  1. G. Tzanetakis and P. Cook, "Musical genre classification of audio signals," in IEEE Transactions on Speech and Audio Processing, (July 2002)
  2. Sebastian Bach et al. “On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation”. In: PLOS ONE 10.7 (July 2015)
  3. Pattarawat Chormai et al. “Disentangled Explanations of Neural Network Predictions by Finding Relevant Subspaces”. In: IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)