{"id":"arxiv/2605.05493","url":"https://arxiv.org/abs/2605.05493","tags":["stat.ME","cond-mat.stat-mech","cs.LG","math.ST"],"score":0,"title":"A renormalization-group inspired lattice-based framework for piecewise generalized linear models","domain":"stat.ME","source":"arxiv","authors":["Joshua C. Chang"],"reviews":0,"subject":"stat.ME","summary":"We formally introduce a class of models inspired by renormalization group (RG) theory, built on additive hierarchical expansions analogous to those appearing in functional ANOVA and mixed-effects models. Like ReLU convolutional neural networks, they are almost everywhere locally linear; unlike ReLU networks, their partition structure is explicit, interpretable, and easy to modify or constrain. In these models, one defines a multidimensional lattice partition of the input space and uses it to scaffold variations in regression parameters. Each dimension of the lattice corresponds to an…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-05-06","reactions":{}}