Laplacian Regularized Few-Shot Learning

We propose a transductive Laplacian-regularized inference for few-shot tasks. Given any feature embedding learned from the base classes, we minimize a quadratic binary-assignment function containing two terms: (1) a unary term assign- ing query samples to the nearest class prototype, and (2) a pairwise Laplacian term encouraging nearby query samples to have consistent label as- signments... Our transductive inference does not re-train the base model, and can be viewed as a graph clustering of the query set, subject to super- vision constraints from the support set. We derive a computationally efficient bound optimizer of a relaxation of our function, which computes inde- pendent (parallel) updates for each query sample, while guaranteeing convergence. Following a sim- ple cross-entropy training on the base classes, and without complex meta-learning strategies, we con- ducted comprehensive experiments over five few- shot learning benchmarks. Our LaplacianShot consistently outperforms state-of-the-art methods by significant margins across different models, settings, and data sets. Furthermore, our trans- ductive inference is very fast, with computational times that are close to inductive inference, and can be used for large-scale few-shot tasks. read more

PDF Abstract ICML 2020 PDF
Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Few-Shot Image Classification CUB 200 5-way 1-shot LaplacianShot Accuracy 80.96 # 8
Few-Shot Image Classification CUB 200 5-way 5-shot LaplacianShot Accuracy 88.68 # 11
Few-Shot Image Classification iNaturalist (227-way multi-shot) LaplacianShot Accuracy 74.97 # 1
Few-Shot Image Classification Mini-Imagenet 5-way (1-shot) LaplacianShot Accuracy 75.57 # 11
Few-Shot Image Classification Mini-Imagenet 5-way (5-shot) LaplacianShot Accuracy 84.72 # 8
Few-Shot Image Classification miniImagenet โ†’ CUB (5-way 1-shot) LaplacianShot Accuracy 55.46 # 1
Few-Shot Image Classification miniImagenet โ†’ CUB (5-way 5-shot) LaplacianShot Accuracy 66.33 # 1
Few-Shot Image Classification Mini-ImageNet-CUB 5-way (5-shot) LaplacianShot Accuracy 66.33 # 3
Few-Shot Image Classification Tiered ImageNet 5-way (1-shot) LaplacianShot Accuracy 80.30 # 5
Few-Shot Image Classification Tiered ImageNet 5-way (5-shot) LaplacianShot Accuracy 87.93 # 6

Methods


No methods listed for this paper. Add relevant methods here