{"id":"arxiv/2602.02593","url":"https://arxiv.org/abs/2602.02593","tags":["cs.LG","cs.AI","math.OC"],"score":0,"title":"Effective Frontiers: A Unification of Neural Scaling Laws","domain":"cs.LG","source":"arxiv","authors":["Jiaxuan Zou","Zixuan Gong","Ye Su","Huayi Tang","Yong Liu"],"reviews":0,"subject":"cs.LG","summary":"Neural scaling laws govern the prediction power-law improvement of test loss with respect to model capacity ($N$), datasize ($D$), and compute ($C$). However, existing theoretical explanations often rely on specific architectures or complex kernel methods, lacking intuitive universality. In this paper, we propose a unified framework that abstracts general learning tasks as the progressive coverage of patterns from a long-tail (Zipfian) distribution. We introduce the Effective Frontier ($k_\\star$), a threshold in the pattern rank space that separates learned knowledge from the unlearned tail.…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-02-01","reactions":{}}