Learning From One And Only One Shot

Abstract

Humans can generalize from only a few examples and from little pretraining on similar tasks. Yet, machine learning (ML) typically requires large data to learn or pre-learn to transfer. Motivated by nativism and artificial general intelligence, we directly model human-innate priors in abstract visual tasks such as character and doodle recognition. This yields a white-box model that learns general-appearance similarity by mimicking how humans naturally ``distort'' an object at first sight. Using just nearest-neighbor classification on this cognitively-inspired similarity space, we achieve human-level recognition with only \(1\)--\(10\) examples per class and no pretraining. This differs from few-shot learning that uses massive pretraining. In the tiny-data regime of MNIST, EMNIST, Omniglot, and QuickDraw benchmarks, we outperform both modern neural networks and classical ML. For unsupervised learning, by learning the non-Euclidean, general-appearance similarity space in a \(k\)-means sty

Stats

• citations4
• S2 citations—
• github stars0
• HF likes0
• heat score5.24
• arxiv keyyu2022learning

Related papers

• Multi-level Similarity Learning For Low-shot Recognition (2019)0.00
• Simpleshot: Revisiting Nearest-neighbor Classification For Few-shot Learning (2019)0.00
• Object-level Representation Learning For Few-shot Image Classification (2018)0.00
• Unsupervised Feature Learning Via Non-parametric Instance-level Discrimination (2018)25.66
• Multimodal Prototypical Networks For Few-shot Learning (2020)14.73
• The Curious Layperson: Fine-grained Image Recognition Without Expert Labels (2021)9.99
• Beginning With You: Perceptual-initialization Improves Vision-language Representation And Alignment (2025)0.00
• Words Are All You Need? Language As An Approximation For Human Similarity Judgments (2022)0.00