LGAIFeb 10, 2023

Project and Probe: Sample-Efficient Domain Adaptation by Interpolating Orthogonal Features

CMU
arXiv:2302.05441v213 citationsh-index: 166
AI Analysis

This addresses sample-efficient domain adaptation for scenarios where target data labels are expensive, offering a lightweight solution with concrete gains.

The paper tackles the problem of adapting pre-trained models to distribution shifts with very limited labeled target data by proposing Project and Probe (Pro^2), which learns orthogonal features and interpolates them, resulting in 5-15% performance improvements over prior methods in experiments across four datasets.

Transfer learning with a small amount of target data is an effective and common approach to adapting a pre-trained model to distribution shifts. In some situations, target data labels may be expensive to obtain, so we may only have access to a limited number of target data points. To make the most of a very small target dataset, we propose a lightweight, sample-efficient approach that learns a diverse set of features and adapts to a target distribution by interpolating these features. Our approach, Project and Probe (Pro$^2$), first learns a linear projection that maps a pre-trained embedding onto orthogonal directions while being predictive of labels in the source dataset. The goal of this step is to learn a variety of predictive features, so that at least some of them remain useful after distribution shift. Pro$^2$ then learns a linear classifier on top of these projected features using a small target dataset. Theoretically, we find that Pro$^2$ results in more sample-efficient generalization by inducing a favorable bias-variance tradeoff. Our experiments on four datasets, with multiple distribution shift settings for each, show that Pro$^2$ improves performance by 5-15% when given limited target data compared to prior methods such as standard linear probing.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes