A Bayesian Interpretation of Adaptive Low-Rank Adaptation

2024/09/16 by Haolin Chen, Chen, Haolin, Philip N. Garner +1 · 1 citation
Decision Sciences · Earth and Planetary Sciences · Physics and Astronomy · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Meteorological Phenomena and Simulations #Model Reduction and Neural Networks #Stock Market Forecasting Methods

paper · pdf · doi:10.48550/arxiv.2409.10673

openalex publication_date 2024/09/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Motivated by the sensitivity-based importance score of the adaptive low-rank adaptation (AdaLoRA), we utilize more theoretically supported metrics, including the signal-to-noise ratio (SNR), along with the Improved Variational Online Newton (IVON) optimizer, for adaptive parameter budget allocation. The resulting Bayesian counterpart not only has matched or surpassed the performance of using the sensitivity-based importance metric but is also a faster alternative to AdaLoRA with Adam. Our theoretical analysis reveals a significant connection between the two metrics, providing a Bayesian perspective on the efficacy of sensitivity as an importance score. Furthermore, our findings suggest that the magnitude, rather than the variance, is the primary indicator of the importance of parameters.

Cited by

• Improving LoRA with Variational Learning

Related

• Learning Adapter Rank via Symmetry Breaking
• Statistical Inference for Rank Allocation in Low-Rank Adaptation
• Adaptively Optimised Adaptive Importance Samplers
• LoFT: Low-Rank Adaptation That Behaves Like Full Fine-Tuning
• Variational Low-Rank Adaptation Using IVON