In multi-label neural collapse, terminal geometry is controlled by the centered label covariance spectrum κ_m derived from label distribution moments, with higher-multiplicity prototypes following class-frequency-weighted synthesis instead of uniform averaging.
Proceedings of the National Academy of Sciences , volume=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.LG 3years
2026 3representative citing papers
SDE-based analysis of multi-task RRG gradient dynamics reveals a double dilemma, addressed by the backbone-agnostic CAME-Grad optimizer that improves clinical metrics by 2.3% on MIMIC-CXR and 1.9% on IU X-Ray across eight methods.
citing papers explorer
-
How Label Imbalance Shapes Geometry: A General Spectral Analysis of Multi-Label Neural Collapse
In multi-label neural collapse, terminal geometry is controlled by the centered label covariance spectrum κ_m derived from label distribution moments, with higher-multiplicity prototypes following class-frequency-weighted synthesis instead of uniform averaging.
-
The Double Dilemma in Multi-Task Radiology Report Generation: A Gradient Dynamics Analysis and Solution
SDE-based analysis of multi-task RRG gradient dynamics reveals a double dilemma, addressed by the backbone-agnostic CAME-Grad optimizer that improves clinical metrics by 2.3% on MIMIC-CXR and 1.9% on IU X-Ray across eight methods.
- Covariance Structure and Coordinate Heterogeneity Govern Binary Quantization of Contrastive Embeddings