ImitAL: Learning Active Learning Strategies from Synthetic Data
This work addresses the problem of efficient and generalizable active learning for large-scale annotation projects, though it appears incremental as it builds on existing AL and imitation learning concepts.
The authors tackled the challenge of active learning (AL) query strategies being domain-specific and computationally expensive by proposing ImitAL, a novel strategy that encodes AL as a learning-to-rank problem using imitation learning trained on synthetic data, achieving superior performance on 15 datasets and improved runtime efficiency compared to 10 state-of-the-art methods.
One of the biggest challenges that complicates applied supervised machine learning is the need for huge amounts of labeled data. Active Learning (AL) is a well-known standard method for efficiently obtaining labeled data by first labeling the samples that contain the most information based on a query strategy. Although many methods for query strategies have been proposed in the past, no clear superior method that works well in general for all domains has been found yet. Additionally, many strategies are computationally expensive which further hinders the widespread use of AL for large-scale annotation projects. We, therefore, propose ImitAL, a novel query strategy, which encodes AL as a learning-to-rank problem. For training the underlying neural network we chose Imitation Learning. The required demonstrative expert experience for training is generated from purely synthetic data. To show the general and superior applicability of \ImitAL{}, we perform an extensive evaluation comparing our strategy on 15 different datasets, from a wide range of domains, with 10 different state-of-the-art query strategies. We also show that our approach is more runtime performant than most other strategies, especially on very large datasets.