---
title: "From Neural Collapse to Label-Limited Evolving Streams: Geometry-Constrained Learning Under Dynamic Class Imbalance"
authors:
  - "Hongliang Wang"
  - "Hongyuan Liu"
  - "Qirui Hao"
  - "Yun Sing Koh"
  - "Qinli Yang"
  - "Junming Shao"
doi: "10.24963/ijcai.2026/341"
url: "https://ijcai.org/proceedings/2026/341"
pdf: "https://ijcai.org/proceedings/2026/0341.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "3070-3078"
track: "Main Track"
keywords:
  - "Machine Learning: Active learning"
  - "Data Mining: Class imbalance and unequal cost"
  - "Data Mining: Mining data streams"
  - "Machine Learning: Semi-supervised learning"
---

# From Neural Collapse to Label-Limited Evolving Streams: Geometry-Constrained Learning Under Dynamic Class Imbalance

Hongliang Wang, Hongyuan Liu, Qirui Hao, Yun Sing Koh, Qinli Yang, Junming Shao

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 3070-3078. DOI: [10.24963/ijcai.2026/341](https://doi.org/10.24963/ijcai.2026/341)

## Abstract

Learning from label-limited streams presents significant challenges, particularly when coupled with concept drift and dynamic class imbalance. Existing works often struggle to maintain a discriminative feature space under these constraints, biasing decision boundaries toward majority classes or outdated concepts. To address this, we propose a novel framework named Neural collapse Inspired Label-limited Evolving stream learning (NILE). Instead of utilizing learnable classifiers, NILE exploits Neural Collapse (NC) geometry to explicitly construct a Simplex Equiangular Tight Frame (ETF) as a fixed classifier, ensuring maximal inter-class separability to guide feature discriminability. To maintain this discrimination with limited labels, NILE employs hybrid active learning to prioritize uncertain and minority samples, and an NC-adapted semi-supervised mechanism to enhance representation learning. NILE further updates the classifier by dynamically adjusting the ETF structure, enabling continual adaptation to drifting concepts. Extensive experiments show that NILE can effectively guide features toward the NC state, significantly outperforming state-of-the-art baselines in nonstationary environments.