Enhancing Transfer Learning with Flexible Nonparametric Posterior Sampling
This work addresses distribution shift issues in transfer learning for machine learning practitioners, but it appears incremental as it builds on existing nonparametric learning methods.
The paper tackled the problem of distribution shifts in transfer learning by introducing nonparametric transfer learning (NPTL), a flexible posterior sampling method, and demonstrated that it surpasses other baselines in Bayesian model averaging performance.
Transfer learning has recently shown significant performance across various tasks involving deep neural networks. In these transfer learning scenarios, the prior distribution for downstream data becomes crucial in Bayesian model averaging (BMA). While previous works proposed the prior over the neural network parameters centered around the pre-trained solution, such strategies have limitations when dealing with distribution shifts between upstream and downstream data. This paper introduces nonparametric transfer learning (NPTL), a flexible posterior sampling method to address the distribution shift issue within the context of nonparametric learning. The nonparametric learning (NPL) method is a recent approach that employs a nonparametric prior for posterior sampling, efficiently accounting for model misspecification scenarios, which is suitable for transfer learning scenarios that may involve the distribution shift between upstream and downstream tasks. Through extensive empirical validations, we demonstrate that our approach surpasses other baselines in BMA performance.