Harnessing Orthogonality to Train Low-Rank Neural Networks
This work addresses training efficiency for neural network practitioners, but it appears incremental as it builds on known orthogonality concepts.
The study tackled the problem of inefficient neural network training by analyzing weight orthogonality, resulting in the OIALR method that integrates into existing workflows with minimal accuracy loss and can surpass conventional setups.
This study explores the learning dynamics of neural networks by analyzing the singular value decomposition (SVD) of their weights throughout training. Our investigation reveals that an orthogonal basis within each multidimensional weight's SVD representation stabilizes during training. Building upon this, we introduce Orthogonality-Informed Adaptive Low-Rank (OIALR) training, a novel training method exploiting the intrinsic orthogonality of neural networks. OIALR seamlessly integrates into existing training workflows with minimal accuracy loss, as demonstrated by benchmarking on various datasets and well-established network architectures. With appropriate hyperparameter tuning, OIALR can surpass conventional training setups, including those of state-of-the-art models.