Neural Fine-Tuning Search for Few-Shot Learning
This addresses the challenge of efficiently adapting pre-trained models to new tasks with limited data, which is crucial for real-world applications where labeled data is scarce.
The paper tackles the problem of designing optimal adaptation strategies for few-shot learning by using neural architecture search to discover the best arrangement of adapters and fine-tuning layers in pre-trained networks, achieving state-of-the-art performance on Meta-Dataset and Meta-Album benchmarks.
In few-shot recognition, a classifier that has been trained on one set of classes is required to rapidly adapt and generalize to a disjoint, novel set of classes. To that end, recent studies have shown the efficacy of fine-tuning with carefully crafted adaptation architectures. However this raises the question of: How can one design the optimal adaptation strategy? In this paper, we study this question through the lens of neural architecture search (NAS). Given a pre-trained neural network, our algorithm discovers the optimal arrangement of adapters, which layers to keep frozen and which to fine-tune. We demonstrate the generality of our NAS method by applying it to both residual networks and vision transformers and report state-of-the-art performance on Meta-Dataset and Meta-Album.