Yang Jiang

$ε$-rank and the Staircase Phenomenon: New Insights into Neural Network Training Dynamics

2024 · arXiv (Cornell University)

Understanding the training dynamics of deep neural networks (DNNs), particularly how they evolve low-dimensional features from high-dimensional data, remains a central challenge in deep learning theory. In this work, we introduce the concept of $ε$-rank, …

Bridging Global Pretraining and Similarity-Based Local Fine-Tuning in GAIN for Imputing Sparse Learner Performance Data

2025

Learner performance data (e.g., correct or incorrect responses) from the interaction logs of Intelligent Tutoring Systems (ITSs) are often sparse, hindering accurate predictions of learner performance and the delivery of effective, adaptive feedback. To address …