Rethinking Few-Shot Adaptation of Vision-Language Models in Two Stages
By: Matteo Farina , Massimiliano Mancini , Giovanni Iacca and more
Potential Business Impact:
Teaches computers to learn new things faster.
An old-school recipe for training a classifier is to (i) learn a good feature extractor and (ii) optimize a linear layer atop. When only a handful of samples are available per category, as in Few-Shot Adaptation (FSA), data are insufficient to fit a large number of parameters, rendering the above impractical. This is especially true with large pre-trained Vision-Language Models (VLMs), which motivated successful research at the intersection of Parameter-Efficient Fine-tuning (PEFT) and FSA. In this work, we start by analyzing the learning dynamics of PEFT techniques when trained on few-shot data from only a subset of categories, referred to as the ``base'' classes. We show that such dynamics naturally splits into two distinct phases: (i) task-level feature extraction and (ii) specialization to the available concepts. To accommodate this dynamic, we then depart from prompt- or adapter-based methods and tackle FSA differently. Specifically, given a fixed computational budget, we split it to (i) learn a task-specific feature extractor via PEFT and (ii) train a linear classifier on top. We call this scheme Two-Stage Few-Shot Adaptation (2SFS). Differently from established methods, our scheme enables a novel form of selective inference at a category level, i.e., at test time, only novel categories are embedded by the adapted text encoder, while embeddings of base categories are available within the classifier. Results with fixed hyperparameters across two settings, three backbones, and eleven datasets, show that 2SFS matches or surpasses the state-of-the-art, while established methods degrade significantly across settings.
Similar Papers
Auxiliary Descriptive Knowledge for Few-Shot Adaptation of Vision-Language Model
CV and Pattern Recognition
Helps AI understand new things better with extra words.
Robust Few-Shot Vision-Language Model Adaptation
CV and Pattern Recognition
Helps AI understand images it hasn't seen.
Complementary Subspace Low-Rank Adaptation of Vision-Language Models for Few-Shot Classification
CV and Pattern Recognition
Helps computers learn new things without forgetting old ones.