Structured Prediction for Conditional Meta-Learning
arXiv:2002.08799
Abstract
The goal of optimization-based meta-learning is to find a single initialization shared across a distribution of tasks to speed up the process of learning new tasks. Conditional meta-learning seeks task-specific initialization to better capture complex task distributions and improve performance. However, many existing conditional methods are difficult to generalize and lack theoretical guarantees. In this work, we propose a new perspective on conditional meta-learning via structured prediction. We derive task-adaptive structured meta-learning (TASML), a principled framework that yields task-specific objective functions by weighing meta-training data on target tasks. Our non-parametric approach is model-agnostic and can be combined with existing meta-learning methods to achieve conditioning. Empirically, we show that TASML improves the performance of existing meta-learning models, and outperforms the state-of-the-art on benchmark datasets.
arXiv comment25 pages, 4 figures, 7 tables
References in corpus (12)
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- Prototypical Networks for Few-shot Learning
- Meta-SGD: Learning to Learn Quickly for Few-Shot Learning
- On First-Order Meta-Learning Algorithms
- A New Approach to Collaborative Filtering: Operator Estimation with Spectral Regularization
- Meta-Learning with Implicit Gradients
- Learning to Optimize
- Hierarchically Structured Meta-learning
- Multimodal Model-Agnostic Meta-Learning via Task-Aware Modulation
- Differential Properties of Sinkhorn Approximation for Learning with Wasserstein Distance
- Multiclass Learning with Simplex Coding
- Weighted Meta-Learning