MMODELYST
Papers/MIC: Maximizing Informational Capacity in Adaptive Representations via Isotropic Subspace Alignment
PAP

MIC: Maximizing Informational Capacity in Adaptive Representations via Isotropic Subspace Alignment

May 28, 2026

arXiv
Abstract

Although multi-scales representation learning enables elastic-dimension embeddings, nested subspaces often suffer from dimensional redundancy and spectral collapse. To address this, we introduce MIC, a framework that optimizes the geometric landscape of multi-granular embeddings through isotropic subspace alignment. MIC employs Soft Collapse Regularization (SCR) to mitigate redundancy between prefix and residual subspaces via cross-correlation penalties, alongside Spectral Isotropy Regularization (SIR) to ensure hyper-spherical uniformity in low-dimensional prefixes. By unifying these strategies through a self-distillation objective, MIC generates semantically dense representations that maintain high discriminative power. Our experiments demonstrate that MIC significantly outperforms standard baselines, particularly in high-compression scenarios where maintaining informational capacity is most critical.

Select text to highlight · click a highlight to remove · saved in this browser only
Authors
Dang Hong Nguyen, Nhi Ngoc-Yen Nguyen, Huy-Hieu Pham
Your notes (browser-local)
saved
arXiv:2605.29987