Reliability and Interpretability in Science and Deep Learning

Abstract

In recent years, the question of the reliability of Machine Learning (ML) methods has acquired significant importance, and the analysis of the associated uncertainties has motivated a growing amount of research. However, most of these studies have applied standard error analysis to ML models, and in particular Deep Neural Network (DNN) models, which represent a rather significant departure from standard scientific modelling. It is therefore necessary to integrate the standard error analysis with a deeper epistemological analysis of the possible differences between DNN models and standard scientific modelling and the possible implications of these differences in the assessment of reliability. This article offers several contributions. First, it emphasises the ubiquitous role of model assumptions (both in ML and traditional Science) against the illusion of theory-free science. Secondly, model assumptions are analysed from the point of view of their (epistemic) complexity, which is shown to be language-independent. It is argued that the high epistemic complexity of DNN models hinders the estimate of their reliability and also their prospect of long-term progress. Some potential ways forward are suggested. Thirdly, this article identifies the close relation between a model's epistemic complexity and its interpretability, as introduced in the context of responsible AI. This clarifies in which sense, and to what extent, the lack of understanding of a model (black-box problem) impacts its interpretability in a way that is independent of individual skills. It also clarifies how interpretability is a precondition for assessing the reliability of any model, which cannot be based on statistical analysis alone. This article focuses on the comparison between traditional scientific models and DNN models. But, Random Forest and Logistic Regression models are also briefly considered.

To appear in Minds and Machines

References in corpus (16)

• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• Aleatoric and Epistemic Uncertainty in Machine Learning: An Introduction to Concepts and Methods
• Neural Tangent Kernel: Convergence and Generalization in Neural Networks
• Hands-on Bayesian Neural Networks -- a Tutorial for Deep Learning Users
• Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
• Evidential Deep Learning to Quantify Classification Uncertainty
• Time Series Data Augmentation for Deep Learning: A Survey
• Predictive Uncertainty Estimation via Prior Networks
• Adversarial Examples Are Not Easily Detected: Bypassing Ten Detection Methods
• Computing Nonvacuous Generalization Bounds for Deep (Stochastic) Neural Networks with Many More Parameters than Training Data
• Flow-based generative models for Markov chain Monte Carlo in lattice field theory
• A Review of Methods for Estimating Algorithmic Complexity: Options, Challenges, and New Directions
• Noise or Signal: The Role of Image Backgrounds in Object Recognition
• Direct Uncertainty Prediction for Medical Second Opinions
• Deep Learning Opacity in Scientific Discovery
• On the role of simplicity in science

Cited by in corpus (3)

• Explainable autoencoder for neutron star dense matter parameter estimation
• Trustworthy AI-based crack-tip segmentation using domain-guided explanations
• High-throughput spin-bath characterization of spin-defects in semiconductors