• Corpus ID: 235458538

Gone Fishing: Neural Active Learning with Fisher Embeddings

  title={Gone Fishing: Neural Active Learning with Fisher Embeddings},
  author={Jordan T. Ash and Surbhi Goel and Akshay Krishnamurthy and Sham M. Kakade},
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… 

Test Distribution-Aware Active Learning: A Principled Approach Against Distribution Shift and Outliers

It is argued that conventional model-based methods for active learning—like BALD—have a fundamental shortfall: they fail to directly account for the testtime distribution of the input variables, and an acquisition strategy is revisited based on maximizing the expected information gained about possible future predictions.

Anti-Concentrated Confidence Bonuses for Scalable Exploration

A practical variant for deep reinforcement learning that is competitive with contemporary intrinsic reward heuristics on Atari benchmarks is developed, using an ensemble of regressors trained to predict random noise from policy network-derived features.

A Framework and Benchmark for Deep Batch Active Learning for Regression

An open-source benchmark with 15 large tabular data sets is introduced, which is used to compare different BMDAL methods and shows that a combination of the novel components yields new state-of-the-art results in terms of RMSE and is computationally efficient.

GALAXY: Graph-based Active Learning at the Extreme

GALAXY automatically and adaptively selects more class-balanced examples for labeling than most other methods for active learning, and its theory and methods demonstrate its superiority over existing state-of-art deep active learning algorithms in unbalanced vision classification settings generated from popular datasets.

Stochastic Batch Acquisition for Deep Active Learning

It is concluded that there is no reason to use top-K batch acquisition in practice, and a stochastic strategy for adapting wellknown acquisition functions to allow batch active learning is provided.

Unifying Approaches in Data Subset Selection via Fisher Information and Information-Theoretic Quantities

The Fisher information is revisited and used to show how several otherwise disparate methods are connected as approximations of information-theoretic quantities.

Neural Design for Genetic Perturbation Experiments

This work provides a theoretically sound framework for iteratively exploring the space of perturbations in pooled batches in order to maximize a target phenotype under an experimental budget and introduces the Optimistic Arm Elimination principle.

Multi-Domain Active Learning: Literature Review and Comparative Study

This work constructs a pipeline of MDAL and presents a comprehensive comparative study of thirty different algorithms, which are established by combining six representative MDL models and commonly used AL strategies, and qualitatively analyze the behaviors of the well-performed strategies and models.

Active Learning in Bayesian Neural Networks with Balanced Entropy Learning Principle

This paper designs and proposes a new uncertainty measure, Balanced Entropy Acquisition (BalEntAcq), which captures the information balance between the uncertainty of underlying softmax probability and the label variable, and demonstrates that it consistently outperforms well-known linearly scalable active learning methods.



Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Bounds

This work designs a new algorithm for batch active learning with deep neural network models that samples groups of points that are disparate and high-magnitude when represented in a hallucinated gradient space, and shows that while other approaches sometimes succeed for particular batch sizes or architectures, BADGE consistently performs as well or better, making it a versatile option for practical active learning problems.

Deep Active Learning over the Long Tail

A novel active learning algorithm that queries consecutive points from the pool using farthest-first traversals in the space of neural activation over a representation layer shows consistent and overwhelming improvement in sample complexity over passive learning (random sampling) for three datasets: MNIST, CIFar-10, and CIFAR-100.

Active Learning for Convolutional Neural Networks: A Core-Set Approach

This work defines the problem of active learning as core-set selection as choosing set of points such that a model learned over the selected subset is competitive for the remaining data points, and presents a theoretical result characterizing the performance of any selected subset using the geometry of the datapoints.

Adversarial Active Learning for Deep Networks: a Margin Based Approach

It is demonstrated empirically that adversarial active queries yield faster convergence of CNNs trained on MNIST, the Shoe-Bag and the Quick-Draw datasets.

Agnostic active learning

The first active learning algorithm which works in the presence of arbitrary forms of noise is state and analyzed, and it is shown that A2 achieves an exponential improvement over the usual sample complexity of supervised learning.

On Warm-Starting Neural Network Training

A closer look is taken at this empirical phenomenon, warm-starting neural network training, which seems to yield poorer generalization performance than models that have fresh random initializations, even though the final training losses are similar.

BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learning

We develop BatchBALD, a tractable approximation to the mutual information between a batch of points and model parameters, which we use as an acquisition function to select multiple informative points

Asymptotic Analysis of Objectives Based on Fisher Information in Active Learning

This paper shows that FIR can be asymptotically viewed as an upper bound of the expected variance of the log-likelihood ratio, and suggests a unifying framework that not only enables to make theoretical comparisons among the existing querying methods based on FIR, but also allows to give insight into the development of new active learning approaches.

Active Deep Learning with Fisher Information for Patch-Wise Semantic Segmentation

A novel diversified AL based on Fisher information (FI) for the first time for CNNs, where gradient computations from backpropagation are used for efficient computation of FI on the large CNN parameter space to achieve accuracy higher than entropy-based querying in transfer learning.

Deep Bayesian Active Learning with Image Data

This paper develops an active learning framework for high dimensional data, a task which has been extremely challenging so far, with very sparse existing literature, and demonstrates its active learning techniques with image data, obtaining a significant improvement on existing active learning approaches.