Artificial Intelligence and Statistics

Abstract

Artificial intelligence (AI) is intrinsically data-driven. It calls for the application of statistical concepts through human-machine collaboration during generation of data, development of algorithms, and evaluation of results. This paper discusses how such human-machine collaboration can be approached through the statistical concepts of population, question of interest, representativeness of training data, and scrutiny of results (PQRS). The PQRS workflow provides a conceptual framework for integrating statistical ideas with human input into AI products and research. These ideas include experimental design principles of randomization and local control as well as the principle of stability to gain reproducibility and interpretability of algorithms and data results. We discuss the use of these principles in the contexts of self-driving cars, automated medical diagnoses, and examples from the authors' collaborative research.

Cited by in corpus (24)

• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• A Gentle Introduction to Deep Learning for Graphs
• Uncertainty Quantification in Machine Learning for Engineering Design and Health Prognostics: A Tutorial
• Macroscopic Traffic Flow Modeling with Physics Regularized Gaussian Process: A New Insight into Machine Learning Applications
• Kernel Flows: from learning kernels from data into the abyss
• Physics discovery in nanoplasmonic systems via autonomous experiments in Scanning Transmission Electron Microscopy
• Energy-entropy competition and the effectiveness of stochastic gradient descent in machine learning
• Exponential data encoding for quantum supervised learning
• MultiScale Probability Map guided Index Pooling with Attention-based learning for Road and Building Segmentation
• Deep regularization and direct training of the inner layers of Neural Networks with Kernel Flows
• Predictive Online Convex Optimization
• GFlowNets for AI-Driven Scientific Discovery
• Density estimation with Gaussian processes for gravitational-wave posteriors
• Statistical Inference for the Population Landscape via Moment Adjusted Stochastic Gradients
• Algorithmic Differentiation for Automated Modeling of Machine Learned Force Fields
• Mirror Descent Search and its Acceleration
• Domain-knowledge Inspired Pseudo Supervision (DIPS) for Unsupervised Image-to-Image Translation Models to Support Cross-Domain Classification
• On the expected behaviour of noise regularised deep neural networks as Gaussian processes
• Data-Efficient Discovery of Hyperelastic TPMS Metamaterials with Extreme Energy Dissipation
• Non-Myopic Multifidelity Bayesian Optimization
• Partial Correlation Graphical LASSO
• BONSAI: Structure-exploiting robust Bayesian optimization for networked black-box systems under uncertainty
• Small errors in random zeroth-order optimization are imaginary
• nuGPR: GPU-Accelerated Gaussian Process Regression with Iterative Algorithms and Low-Rank Approximations