Amortized Bayesian Meta-Learning for Low-Rank Adaptation of Large Language Models
This addresses computational efficiency and generalization issues for researchers and practitioners fine-tuning LLMs, though it appears incremental as it adapts amortized Bayesian meta-learning from smaller models to LLMs.
The paper tackles the problem of expensive memory and computation costs in fine-tuning large language models with low-rank adaptation while improving generalization to unseen datasets, proposing ABMLL which outperforms existing methods on Unified-QA and CrossFit benchmarks in accuracy and expected calibration error.
Fine-tuning large language models (LLMs) with low-rank adaptaion (LoRA) is a cost-effective way to incorporate information from a specific dataset. However, it is often unclear how well the fine-tuned LLM will generalize, i.e., how well it will perform on unseen datasets. Methods have been proposed to improve generalization by optimizing with in-context prompts, or by using meta-learning to fine-tune LLMs. However, these methods are expensive in memory and computation, requiring either long-context prompts or saving copies of parameters and using second-order gradient updates. To address these challenges, we propose Amortized Bayesian Meta-Learning for LoRA (ABMLL). This method builds on amortized Bayesian meta-learning for smaller models, adapting this approach to LLMs while maintaining its computational efficiency. We reframe task-specific and global parameters in the context of LoRA and use a set of new hyperparameters to balance reconstruction accuracy and the fidelity of task-specific parameters to the global ones. ABMLL provides effective generalization and scales to large models such as Llama3-8B. Furthermore, as a result of using a Bayesian framework, ABMLL provides improved uncertainty quantification. We test ABMLL on Unified-QA and CrossFit datasets and find that it outperforms existing methods on these benchmarks in terms of both accuracy and expected calibration error.