Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

We propose an algorithm for meta-learning that is model-agnostic, in the sense that it is compatible with any model trained with gradient descent and applicable to a variety of different learning problems, including classification, regression, and reinforcement learning. The goal of meta-learning is to train a model on a variety of learning tasks, such that it can solve new learning tasks using only a small number of training samples... (read more)

PDF Abstract ICML 2017 PDF ICML 2017 Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK BENCHMARK
Few-Shot Image Classification OMNIGLOT - 1-Shot, 5-way MAML Accuracy 98.7 # 10
Few-Shot Image Classification OMNIGLOT - 5-Shot, 5-way MAML Accuracy 99.9 # 2

Results from Other Papers


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK SOURCE PAPER COMPARE
Few-Shot Image Classification Mini-Imagenet 5-way (1-shot) MAML Accuracy 48.7 # 55
Few-Shot Image Classification Mini-Imagenet 5-way (5-shot) MAML Accuracy 63.1 # 52
Few-Shot Image Classification Mini-ImageNet-CUB 5-way (1-shot) MAML (Finn et al., 2017) Accuracy 40.15 # 7
Image Classification Tiered ImageNet 5-way (5-shot) MAML+Transduction Accuracy 70.83 # 4
Image Classification Tiered ImageNet 5-way (5-shot) MAML Accuracy 70.30 # 5

Methods used in the Paper


METHOD TYPE
TRPO
Policy Gradient Methods
Linear Layer
Feedforward Networks
Softmax
Output Functions
Batch Normalization
Normalization
Max Pooling
Pooling Operations
ReLU
Activation Functions
Convolution
Convolutions
MAML
Meta-Learning Algorithms