Meta-Learning with Temporal Convolutions

By Nikhil Mishra, Mostafa Rohaninejad, Xi Chen, Pieter Abbeel. 2017

204 citations recorded by the source corpus.

Deep neural networks excel in regimes with large amounts of data, but tend to struggle when data is scarce or when they need to adapt quickly to changes in the task. Recent work in meta-learning seeks to overcome this shortcoming by training a meta-learner on a distribution of similar tasks; the goal is for the meta-learner to generalize to novel but related tasks by learning a high-level strategy that captures the essence of the problem it is asked to solve. However, most recent approaches to meta-learning are extensively hand-designed, either using architectures that are specialized to a particular application, or hard-coding algorithmic components that tell the meta-learner how to solve the task. We propose a class of simple and generic meta-learner architectures, based on temporal convolutions, that is domain- agnostic and has no particular strategy or algorithm encoded into it. We validate our temporal-convolution-based meta-learner (TCML) through experiments pertaining to both supervised and reinforcement learning, and demonstrate that it outperforms state-of-the-art methods that are less general and more complex.

Subfields: Artificial Intelligence, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Meta & Continual Learning

Citations in the corpus

Meta-Learning for Low-Resource Neural Machine Translation (2018, background): “…two categories of meta-learning: 1. learning a meta-policy for updating model parameters (see, e.g., Andrychowicz et al., 2016; Ha et al., 2016a; Mishra et al., 2017) 2. learning a good parameter initialization for fast adaptation (see, e.g., Finn et al., 2017; Vinyals et al., 2016; Snell et…”
Zero-shot Text Classification With Generative Language Models (2019, methodology): “However, parallel work such as Memory Augmented Neural Networks (Santoro et al., 2016) and Simple Neural Attentive Learners (Mishra et al., 2017) demonstrate the effectiveness of architecture based meta-learning.”
Meta-Learning for Semi-Supervised Few-Shot Classification (2018, background): “Meta-learning may take the form of learning a shared metric (Vinyals et al., 2016; Snell et al., 2017), a common initialization for few-shot classiﬁers (Ravi & Larochelle, 2017; Finn et al., 2017) or a generic inference network (Santoro et al., 2016; Mishra et al., 2017).”
Meta-SGD: Learning to Learn Quickly for Few Shot Learning (2017, background): “Such a two-tiered learning to learn strategy for meta-learning has been applied successfully to few-shot learning on classiﬁcation [7, 18, 19, 25], regression [7, 19], and reinforcement learning [6, 7, 17, 23, 26].”
Solving Rubik's Cube with a Robot Hand (2019, methodology): “Algorithms such as MAML [33] and SNAIL [69] have been developed to improve the sample efficiency of reinforcement learning agents.”
Recasting Gradient-Based Meta-Learning as Hierarchical Bayes (2018, methodology): “Improved performance on miniImageNet has been reported by several works (Mishra et al., 2017; Munkhdalai & Yu, 2017; Sung et al., 2017) by making use of a model architecture with significantly more parameters than the methods in Table 1.”
Meta-Learning and Universality: Deep Representations and Gradient Descent can Approximate any Learning Algorithm (2017, methodology): “In the ﬁrst approach (Santoro et al., 2016; Duan et al., 2016; Wang et al., 2016; Mishra et al., 2017), there is a meta-learner model g with parameters φ which takes as input the dataset D T for a particular task T and a new test input x (cid:63) , and outputs the estimated output ˆ y (cid:63) for…”
Dynamic Few-Shot Visual Learning Without Forgetting (2018, background): “[12] instead propose a generic temporal convolutional network that given as input a sequence of a few labeled training examples and then an unlabeled test example, it predicts the label of that test example.”
Learning to Compare: Relation Network for Few-Shot Learning (2017, methodology): “When trained with 5-way 15 query per training episode, [36] only got 46 .”
Few-Shot Learning with Graph Neural Networks (2017, neutral): “The TCML approach from Mishra et al. (2017) is in the same conﬁdence interval for 3 out of 4 experiments, but it is slightly better for the 20-”
Bayesian Model-Agnostic Meta-Learning (2018, background): “…learning datasets (Lake et al., 2015; Ravi & Larochelle, 2017), have rekindled the interest in this problem from the perspective of deep networks for few-shot learning (Santoro et al., 2016; Vinyals et al., 2016; Snell et al., 2017; Duan et al., 2016; Finn et al., 2017; Mishra et al., 2017).”
Learning Rapid-Temporal Adaptations (2017, background): “Recently, there has been a shift towards building flexible models for these problems within the metalearning paradigm [23, 32].”
Online Meta-Learning (2019, neutral): “…2017; Ravi & Larochelle, 2017) by having one model output the weights of another (Ha et al., 2017), while other works have used recurrent models that learn by ingesting datasets directly (San-toro et al., 2016; Duan et al., 2016; Wang et al., 2016; Munkhdalai & Yu, 2017; Mishra et al., 2017).”
Improving Generalization via Scalable Neighborhood Component Analysis (2018, background): “Two previous works [27,39] have reported their performance with similar ResNet18 architectures.”