Technometrics · 1989 · 492 citations · 0 references
Natural SciencesComputer ScienceArtificial IntelligenceStatistical LearningStatisticsBayesian InferenceBayesian StatisticsComputational StatisticsInferential StatisticsStatistical AnalysisStatistical InferenceStatistical Learning TheoryStatistical Simulation
Artificial intelligence relies on data‑driven processes that necessitate statistical concepts such as experimental design, randomization, local control, and stability to ensure reproducibility and interpretability. The paper aims to outline a human‑machine collaboration approach using the PQRS framework—population, question of interest, representativeness, and scrutiny—to integrate statistical ideas into AI. The authors illustrate the PQRS workflow by applying these statistical principles to self‑driving cars, automated medical diagnosis, and their own collaborative projects.
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.