{"id":"arxiv/2602.23326","url":"https://arxiv.org/abs/2602.23326","tags":["math.PR","cond-mat.dis-nn"],"score":0,"title":"Spin Glass Concepts in Computer Science, Statistics, and Learning","domain":"math.PR","source":"arxiv","authors":["Andrea Montanari"],"reviews":0,"subject":"math.PR","summary":"Spin glass theory studies the structure of sublevel sets and minima (or near-minima) of certain classes of random functions in high dimension. Near-minima of random functions also play an important role in high-dimensional statistics and statistical learning, where minimizing the empirical risk (which is a random function of the model parameters) is the method of choice for learning a statistical model from noisy data. Finally, near-minima of random functions are obviously central to average-case analysis of optimization algorithms. Computer science, statistics, and machine learning…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-02-26","reactions":{}}