The goal of this paper is to design image classification systems that, after an initial multi-task training phase, can automatically adapt to new tasks encountered at test time. We introduce a conditional neural process based approach to the multi-task classification setting for this purpose, and establish connections to the meta-and few-shot learning literature. The resulting approach, called CNAPs, comprises a classifier whose parameters are modulated by an adaptation network that takes the current task's dataset as input. We demonstrate that CNAPs achieves state-of-the-art results on the challenging Meta-Dataset benchmark indicating high-quality transfer-learning. We show that the approach is robust, avoiding both over-fitting in low-shot regimes and under-fitting in high-shot regimes. Timing experiments reveal that CNAPs is computationally efficient at test-time as it does not involve gradient based adaptation. Finally, we show that trained models are immediately deployable to continual learning and active learning where they can outperform existing approaches that do not leverage transfer learning.

The goal of this paper is to design image classification systems that, after an initial multi-task training phase, can automatically adapt to new tasks encountered at test time. We introduce a conditional neural process based approach to the multi-task classification setting for this purpose, and establish connections to the meta- and few-shot learning literature. The resulting approach, called CNAPs, comprises a classifier whose parameters are modulated by an adaptation network that takes the current task's dataset as input. We demonstrate that CNAPs achieves state-of-the-art results on the challenging Meta-Dataset benchmark indicating high-quality transfer-learning. We show that the approach is robust, avoiding both over-fitting in low-shot regimes and under-fitting in high-shot regimes.

Quality: The technical content of the paper is well motivated and the approach taken is interesting. However, a few things are worth mentioning. 1 - The classification parameters for a given class are generated independently from the other classes. This means that the classifier is more likely to act as a prototypical model than a discriminative one. 2 - In the adaptation network, the auto-regressive component is not technically motivated. The fact that it improves results just shows the lack of capacity in the FiLM network as a way to modulate the feature extractor parameters alone. Did you compare different ways of modulating the feature extractor parameters? 3 - z_G is computed using only the inputs from the query set, what about the labels? 4 - The statement " Allowing θ to adapt during the second phase violates the principle of "train as you test", i.e., when test tasks are encountered, θ will be fixed, so it is important to simulate this scenario during training " is technically false as within each meta-learning step θ will be fixed even when is not pretrained. Thus, the justification for the training procedure is a bit weak despite the comparison between the proposed approach and the classical one.

This paper proposes a novel few-shot learning method, with a specific application focus to fine-tuning CV object classification models from pre-trained features. Different from previous few-shot learning or CNP work, this work tries to address a convincing real world use case. Its novelties include inference amortization for head models, adaptation of the feature network on each task using a novel autoregressive architecture. One point of improving the paper is to move details about the autoregressive model structure and the adaptation network into the main text. Too many relevant details are just in the supplemental material.

The goal of this paper is to design image classification systems that, after an initial multi-task training phase, can automatically adapt to new tasks encountered at test time. We introduce a conditional neural process based approach to the multi-task classification setting for this purpose, and establish connections to the meta- and few-shot learning literature. The resulting approach, called CNAPs, comprises a classifier whose parameters are modulated by an adaptation network that takes the current task's dataset as input. We demonstrate that CNAPs achieves state-of-the-art results on the challenging Meta-Dataset benchmark indicating high-quality transfer-learning. We show that the approach is robust, avoiding both over-fitting in low-shot regimes and under-fitting in high-shot regimes.

The goal of this paper is to design image classification systems that, after an initial multi-task training phase, can automatically adapt to new tasks encountered at test time. We introduce a conditional neural process based approach to the multi-task classification setting for this purpose, and establish connections to the meta- and few-shot learning literature. The resulting approach, called CNAPs, comprises a classifier whose parameters are modulated by an adaptation network that takes the current task's dataset as input. We demonstrate that CNAPs achieves state-of-the-art results on the challenging Meta-Dataset benchmark indicating high-quality transfer-learning. We show that the approach is robust, avoiding both over-fitting in low-shot regimes and under-fitting in high-shot regimes.