Discovering General-Purpose Active Learning Strategies
arXiv:1810.04114
Abstract
We propose a general-purpose approach to discovering active learning (AL) strategies from data. These strategies are transferable from one domain to another and can be used in conjunction with many machine learning models. To this end, we formalize the annotation process as a Markov decision process, design universal state and action spaces and introduce a new reward function that precisely model the AL objective of minimizing the annotation cost. We seek to find an optimal (non-myopic) AL strategy using reinforcement learning. We evaluate the learned strategies on multiple unrelated domains and show that they consistently outperform state-of-the-art baselines.
References in corpus (8)
- Playing Atari with Deep Reinforcement Learning
- Neural Architecture Search with Reinforcement Learning
- Designing Neural Network Architectures using Reinforcement Learning
- Active One-shot Learning
- Meta-Learning Transferable Active Learning Policies by Deep Reinforcement Learning
- Active Learning from Weak and Strong Labelers
- Learning Algorithms for Active Learning
- A Meta-Learning Approach to One-Step Active Learning
Cited by in corpus (6)
- Reinforced active learning for image segmentation
- Towards Understanding the Behaviors of Optimal Deep Active Learning Algorithms
- Learning active learning at the crossroads? evaluation and discussion
- Active Learning for Argument Mining: A Practical Approach
- ImitAL: Learning Active Learning Strategies from Synthetic Data
- Active Offline Policy Selection