BLoB: Bayesian Low-Rank Adaptation by Backpropagation for Large Language Models
This addresses uncertainty quantification for LLMs in domain-specific applications, representing an incremental improvement over post-training Bayesian methods.
The paper tackles the problem of overconfidence in large language models during inference on domain-specific tasks with limited data by proposing BLoB, an algorithm that jointly adjusts mean and covariance parameters during fine-tuning, resulting in improved generalization and uncertainty estimation on in-distribution and out-of-distribution data.
Large Language Models (LLMs) often suffer from overconfidence during inference, particularly when adapted to downstream domain-specific tasks with limited data. Previous work addresses this issue by employing approximate Bayesian estimation after the LLMs are trained, enabling them to quantify uncertainty. However, such post-training approaches' performance is severely limited by the parameters learned during training. In this paper, we go beyond post-training Bayesianization and propose Bayesian Low-Rank Adaptation by Backpropagation (BLoB), an algorithm that continuously and jointly adjusts both the mean and covariance of LLM parameters throughout the whole fine-tuning process. Our empirical results verify the effectiveness of BLoB in terms of generalization and uncertainty estimation, when evaluated on both in-distribution and out-of-distribution data.