Abstract:Humans can solve image classification tasks by learning from a few images and reusing prior-knowledge. In Artificial Intelligence, deep-learning models have been implemented to simulate human learning and face problems with little data available, few-shot learning. Nevertheless, one crucial problem of deep-learning is the selection of architectures and initial parameters that accomplish the requirements for a specific task. Therefore, we propose a novel methodology based on the meta-learning paradigm, which reuses prior- and meta-knowledge to select the best architecture and its initial parameters to solve different few-shot image classification tasks, as humans do. Our methodology was designed to understand the knowledge flow of the meta-learning paradigm by dividing the learning into prior-models, meta-models, meta-unsupervised algorithm, and base-models. We considered 9 architectures of deep convolutional neural networks as prior-models. Also, we propose a meta-unsupervised algorithm inspired by the human-cognitive problem-solving process, which acquires knowledge by solving tasks to recommend initial parameters to solve other new tasks. Furthermore, we propose a New Task Distribution Scheme to better evaluate few-shot learning models and analyze the difficulty of the new tasks. We evaluated our meta-learning methodology by applying the NTDS to the Mini-ImageNet and Caltech-UCSD-Birds-200-2011 (CUB) datasets. The achieved Adjusted Mutual Information average scores are 0.955 (zero-shot) and 0.957 (five-shot) for Mini-ImageNet, and 0.822 (zero-shot) and 0.848 (five-shot) for CUB. Moreover, we demonstrated the applicability of our meta-learning methodology in a real-world scenario using a new copro-parasite dataset. The performances of our meta-learning methodology are competitive regarding the state-of-the-art unsupervised few-shot learning models.

Peer Review #2 of "research on Image Classification Method Based on Improved Multi-Scale Relational Network (V0.1)"

Peer Review #1 of "research on Image Classification Method Based on Improved Multi-Scale Relational Network (V0.1)"

Research on image classification method based on improved multi-scale relational network

A Few Shot Classification Methods Based on Multiscale Relational Networks

Multi-scale Relation Network for Few-Shot Learning Based on Meta-learning

Evolving parametrized Loss for Image Classification Learning on Small Datasets

ReMarNet: Conjoint Relation and Margin Learning for Small-Sample Image Classification

MetaDelta: A Meta-Learning System for Few-shot Image Classification

RS-MetaNet: Deep meta metric learning for few-shot remote sensing scene classification

Small sample classification of hyperspectral image using model-agnostic meta-learning algorithm and convolutional neural network

Meta-seg: A Survey of Meta-Learning for Image Segmentation

Class Regularization: Improve Few-shot Image Classification by Reducing Meta Shift

Meta Generalized Network for Few-Shot Classification.

Meta-learning methodology based on meta-unsupervised algorithm for meta-model selection to solve few-shot base-tasks

A Joint Optimization Framework of the Embedding Model and Classifier for Meta-Learning

A Two-Stage Spiking Meta-Learning Method for Few-Shot Classification

Differentiable Meta-learning Model for Few-shot Semantic Segmentation

Perceptual quality assessment for no-reference image via optimization-based meta-learning

A meta-learning based progressive prototype network technique in image classification

Multi-distance Metric Network for Few-Shot Learning

Local Mutual Metric Network for Few-Shot Image Classification