# BayesAdapter: Being Bayesian, Inexpensively and Reliably, via Bayesian Fine-tuning

arXiv: https://arxiv.org/abs/2010.01979 · HTML: https://mlsanity.com/paper/2010.01979/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Oct 2020.

## Built on by (most cited first, 8 of 8)

- [Accelerated Linearized Laplace Approximation for Bayesian Deep Learning](https://mlsanity.com/paper/2210.12642/) (2022)
- [Variational Linearized Laplace Approximation for Bayesian Deep Learning](https://mlsanity.com/paper/2302.12565/) (2023)
- [Efficient Detection of LLM-generated Texts with a Bayesian Surrogate Model](https://mlsanity.com/paper/2305.16617/) (2023)
- [Towards Accelerated Model Training via Bayesian Data Selection](https://mlsanity.com/paper/2308.10544/) (2023)
- [Dynamic ensemble selection based on Deep Neural Network Uncertainty Estimation for Adversarial Robustness](https://mlsanity.com/paper/2308.00346/) (2023)
- [Fixed-Mean Gaussian Processes for Post-hoc Bayesian Deep Learning](https://mlsanity.com/paper/2412.04177/) (2024)
- [Deep Ensemble as a Gaussian Process Approximate Posterior](https://mlsanity.com/paper/2205.00163/) (2022)
- [Try to Avoid Attacks: A Federated Data Sanitization Defense for Healthcare IoMT Systems](https://mlsanity.com/paper/2211.01592/) (2022)

## Abstract (arXiv)

Despite their theoretical appealingness, Bayesian neural networks (BNNs) are left behind in real-world adoption, mainly due to persistent concerns on their scalability, accessibility, and reliability. In this work, we develop the BayesAdapter framework to relieve these concerns. In particular, we propose to adapt pre-trained deterministic NNs to be variational BNNs via cost-effective Bayesian fine-tuning. Technically, we develop a modularized implementation for the learning of variational BNNs, and refurbish the generally applicable exemplar reparameterization trick through exemplar parallelization to efficiently reduce the gradient variance in stochastic variational inference. Based on the lightweight Bayesian learning paradigm, we conduct extensive experiments on a variety of benchmarks, and show that our method can consistently induce posteriors with higher quality than competitive baselines, yet significantly reducing training overheads. Code is available at https://github.com/thudzj/ScalableBDL.