Hyunji Jung

2 accepted papers

2026

Mitigating Staleness in Asynchronous Pipeline Parallelism via Basis Rotation

ICML 2026poster

Asynchronous pipeline parallelism maximizes hardware utilization by eliminating the pipeline bubbles inherent in synchronous execution, offering a path toward efficient large-scale distributed training. However, this efficiency gain can be compromised by gradient staleness, where the immediate model…

Cited by 0 Source ↗Scholar ↗

2025

Convergence and Implicit Bias of Gradient Descent on Continual Linear Classification

Hyunji Jung, Hanseul Cho, Chulhee Yun

ICLR 2025poster

We study continual learning on multiple linear classification tasks by sequentially running gradient descent (GD) for a fixed budget of iterations per each given task. When all tasks are jointly linearly separable and are presented in a cyclic/random order, we show the directional convergence of the…

Cited by 2 Source ↗PDF ↗Scholar ↗