Gone Fishing: Neural Active Learning with Fisher Embeddings
This addresses the need for efficient active learning algorithms compatible with neural networks, offering a practical solution for data selection in machine learning applications.
The paper tackles the problem of active learning for deep neural networks by proposing BAIT, a batch active learning algorithm based on Fisher information, which outperforms previous state-of-the-art methods on classification and regression tasks.
There is an increasing need for effective active learning algorithms that are compatible with deep neural networks. This paper motivates and revisits a classic, Fisher-based active selection objective, and proposes BAIT, a practical, tractable, and high-performing algorithm that makes it viable for use with neural models. BAIT draws inspiration from the theoretical analysis of maximum likelihood estimators (MLE) for parametric models. It selects batches of samples by optimizing a bound on the MLE error in terms of the Fisher information, which we show can be implemented efficiently at scale by exploiting linear-algebraic structure especially amenable to execution on modern hardware. Our experiments demonstrate that BAIT outperforms the previous state of the art on both classification and regression problems, and is flexible enough to be used with a variety of model architectures.