Abstract:Neural multivariate regression underpins a wide range of domains, including control, robotics, and finance, yet the geometry of its learned representations remains poorly characterized. While neural collapse has been shown to benefit generalization in classification, we find that analogous collapse in regression consistently degrades performance. To explain this contrast, we analyze regression models through the lens of intrinsic dimension. Across control tasks and synthetic datasets, we estimate the intrinsic dimension of last-layer features (ID_H) and compare it with that of the regression targets (ID_Y). Collapsed models exhibit ID_H < ID_Y, leading to over-compression and poor generalization, whereas non-collapsed models typically maintain ID_H > ID_Y. For the non-collapsed models, performance with respect to ID_H depends on the data quantity and noise levels. From these observations, we identify two regimes (over-compressed and under-compressed) that determine when expanding or reducing feature dimensionality improves performance. Our results provide new geometric insights into neural regression collapse and suggest practical strategies for enhancing generalization.
| Comments: | 36 pages, 21 figures |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2510.01105 [cs.LG] |
| (or arXiv:2510.01105v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2510.01105 arXiv-issued DOI via DataCite |
From: Zixuan Dong [view email]
[v1]
Wed, 1 Oct 2025 16:50:57 UTC (746 KB)
[v2]
Fri, 8 May 2026 13:22:44 UTC (1,696 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。