Expressive and Generalizable Low-rank Adaptation for Large Models via Slow Cascaded Learning

TL;DR

LoRASC is a novel low-rank adaptation method for large models that improves expressiveness and generalization through cascaded learning, slow-fast updates, and noisy tuning, outperforming existing methods across diverse tasks.

Contribution

The paper introduces LoRASC, a new low-rank adaptation technique that enhances expressiveness and robustness of large models via cascaded learning and innovative update mechanisms.

Findings

Significantly outperforms existing LoRA variants on multiple benchmarks.

Reduces overfitting and improves model stability.

Enhances robustness to out-of-distribution data.

Abstract

Efficient fine-tuning plays a fundamental role in modern large models, with low-rank adaptation emerging as a particularly promising approach. However, the existing variants of LoRA are hampered by limited expressiveness, a tendency to overfit, and sensitivity to hyperparameter settings. This paper presents LoRA Slow Cascade Learning (LoRASC), an innovative technique designed to enhance LoRA's expressiveness and generalization capabilities while preserving its training efficiency. Our approach augments expressiveness through a cascaded learning strategy that enables a mixture-of-low-rank adaptation, thereby increasing the model's ability to capture complex patterns. Additionally, we introduce a slow-fast update mechanism and cascading noisy tuning to bolster generalization. The extensive experiments on various language and vision datasets, as well as robustness benchmarks, demonstrate that the proposed method not only significantly outperforms existing baselines, but also mitigates overfitting, enhances model stability, and improves OOD robustness. Code will be release in https://github.com/microsoft/LoRASC very soon.