• Corpus ID: 240070681

Selective Sampling for Online Best-arm Identification

  title={Selective Sampling for Online Best-arm Identification},
  author={Romain Camilleri and Zhihan Xiong and Maryam Fazel and Lalit P. Jain and Kevin G. Jamieson},
This work considers the problem of selective-sampling for best-arm identification. Given a set of potential optionsZ ⊂ R, a learner aims to compute with probability greater than 1 − δ, arg maxz∈Z z>θ∗ where θ∗ is unknown. At each time step, a potential measurement xt ∈ X ⊂ R is drawn IID and the learner can either choose to take the measurement, in which case they observe a noisy measurement of x>θ∗, or to abstain from taking the measurement and wait for a potentially more informative point to… 

Figures from this paper

Achieving Minimax Rates in Pool-Based Batch Active Learning

This paper analyzes a novel stage-wise greedy algorithm and shows that, as a function of the label complexity, the excess risk of this algorithm matches the known minimax rates in standard statistical learning settings.

Fast Rates in Pool-Based Batch Active Learning

This paper theoretically investigates batch active learning in the practically relevant scenario where the unlabeled pool of data is available beforehand ( pool-based active learning) and analyzes a novel stage-wise greedy algorithm and shows that the excess risk of this algorithm matches the known minimax rates in standard statistical learning settings.

Active Learning with Safety Constraints

This work proposes an adaptive experimental design-based algorithm, which it shows efficiently trades off between the difficulty of showing an arm is unsafe vs suboptimal, and results are the first on best-arm identification in linear bandits with safety constraints.

Data Representativity for Machine Learning and AI Systems

This paper analyzes data representativity in scientific literature related to AI and sampling, and proposes a framework of questions for creating and documenting data, with data representative in mind, as an addition to existing datasheets for datasets.



Robust bounds for classification via selective sampling

A new algorithm for binary classification in the selective sampling protocol that approximates the margin of the Bayes optimal classifier to any desired accuracy ε by asking Õ (d/ε2) queries (in the RKHS case d is replaced by a suitable spectral quantity).

Toward a General Theory of Online Selective Sampling: Trading Off Mistakes and Queries

This work explores various properties of the optimal trade-off curve, both abstractly (for general VC classes), and more-concretely for several constructed examples that expose important properties ofThe trade-offs.

On the Complexity of Best-Arm Identification in Multi-Armed Bandit Models

This work introduces generic notions of complexity for the two dominant frameworks considered in the literature: fixed-budget and fixed-confidence settings, and provides the first known distribution-dependent lower bound on the complexity that involves information-theoretic quantities and holds when m ≥ 1 under general assumptions.

Best-Arm Identification in Linear Bandits

The importance of exploiting the global linear structure to improve the estimate of the reward of near-optimal arms is shown and the connection to the G-optimality criterion used in optimal experimental design is pointed out.

Improved Algorithms for Agnostic Pool-based Active Classification

This work proposes an algorithm that, in contrast to uniform sampling over the disagreement region, solves an experimental design problem to determine a distribution over examples from which to request labels and demonstrates that the algorithm is superior to state of the art agnostic active learning algorithms on image classification datasets.

High-Dimensional Experimental Design and Kernel Bandits

This work proposes a rounding procedure that frees N of any dependence on the dimension d, while achieving nearly the same performance guarantees of existing rounding procedures.

A Modern Introduction to Online Learning

This monograph introduces the basic concepts of Online Learning through a modern view of Online Convex Optimization, and presents first-order and second-order algorithms for online learning with convex losses, in Euclidean and non-Euclidean settings.

Theory of Disagreement-Based Active Learning

Recent advances in the understanding of the theoretical benefits of active learning are described, and implications for the design of effective active learning algorithms are described.

Selective sampling algorithms for cost-sensitive multiclass prediction

This paper proposes selective sampling algorithms, which process the data in a streaming fashion, querying only a subset of the labels, and establishes that the gains of active learning over passive learning can range from none to exponentially large, based on a natural notion of margin.

Selective sampling and active learning from single and multiple teachers

This work presents a new online learning algorithm in the selective sampling framework, where labels must be actively queried before they are revealed, and extends the algorithm and analysis to the multiple-teacher setting, where the algorithm can choose which subset of teachers to query for each label.