LSFSL: Leveraging Shape Information in Few-shot Learning
Open Access
- 1 June 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 4971-4980
Abstract
Few-shot learning (FSL) techniques seek to learn the underlying patterns in data using fewer samples, analogous to how humans learn from limited experience. In this limiteddata scenario, the challenges associated with deep neural networks, such as shortcut learning and texture bias behaviors, are further exacerbated. Moreover, the significance of addressing shortcut learning is not yet fully explored in the few-shot setup. To address these issues, we propose LSFSL, which enforces the model to learn more generalizable features utilizing the implicit prior information present in the data. Through comprehensive analyses, we demonstrate that LSFSL-trained models are less vulnerable to alteration in color schemes, statistical correlations, and adversarial perturbations leveraging the global semantics in the data. Our findings highlight the potential of incorporating relevant priors in few-shot approaches to increase robustness and generalization.Keywords
All Related Versions
This publication has 21 references indexed in Scilit:
- Few-Shot Learning With Embedded Class Models and Shot-Free Meta Training2019
- Skeletal descriptions of shape provide unique perceptual information for object recognitionScientific Reports, 2019
- Meta-Learning With Differentiable Convex Optimization2019
- Shape from Contour: Computation and RepresentationAnnual Review of Vision Science, 2018
- Learning to Compare: Relation Network for Few-Shot Learning2018
- Low-Shot Visual Recognition by Shrinking and Hallucinating Features2017
- Arbitrary Style Transfer in Real-Time with Adaptive Instance Normalization2017
- How Specific is the Shape Bias?Child Development, 2003
- A new view of language acquisitionProceedings of the National Academy of Sciences, 2000
- The importance of shape in early lexical learningCognitive Development, 1988