Online Choice of Active Learning Algorithms

By Yoram Baram, Ran El-Yaniv, Kobi Luz. 2003

457 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep Active Learning with a Neural Architecture Search (2018, background): “A common approach for this normalization has already been proposed in [15, 2], and can be deﬁned as follows.”
Active Learning with Feedback on Features and Instances (2006, methodology): “3 Evaluation The deficiency measure was proposed by Baram et al. (2003) as a measure of the speed of an active learning algorithm, useful for comparing different active learning algorithms.”
Online Learning: A Comprehensive Survey (2018, background): “Baram et al. (2004) considers the setting of how to online combine an ensemble of active learners, which is executed based on a maximum entropy criterion.”
Dual Strategy Active Learning (2007, methodology): “We compared the performance of DUAL with that of DWUS, uncertainty sampling, representative sampling(2) [8], density-based sampling and the COMB method of [10].”
Get another label? improving data quality and data mining using multiple, noisy labelers (2008, background): “Active learning (cf., optimal experimental design [33]) uses the existing model to help select additional data for which to acquire labels [1, 14, 23].”
A machine learning approach to sentiment analysis in multilingual Web texts (2009, methodology): “These examples hold little new information, but instead serve the purpose of fine-tuning the current decision boundary of the classifier (Baram et al. 2004).”
A theory of transfer learning with applications to active learning (2012, background): “The goal of transfer learning is to leverage previous learning and experience to more efﬁciently learn novel, but related, concepts, compared to what would be possible without this prior experience.”
The strategic student approach for life-long exploration and learning (2012, methodology): “A third family of approaches have considered internal mechanisms for active choice of either training examples [8]– [10], tasks to be explored [11], learning methods to be used [12], [13] or questions to pose to a teacher [14].”
Active choice of teachers, learning strategies and goals for a socially guided intrinsic motivation learner (2012, background): “For instance, [5] selects among di erent learning strategies depending on the results for di erent outcomes.”
The Multi-fidelity Multi-armed Bandit (2016, background): “Similarly one might consider algorithm selection for machine learning problems [4], where the goal is to be competitive with the best among a set of learning algorithms for a task.”
Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Bounds (2019, background): “For example, Baram et al. (2004); Hsu and Lin (2015) present meta-active learning algorithms that can combine the advantages of different active learning algorithms.”
Active learning and discovery of object categories in the presence of unnameable instances (2015, background): “Finally, combinations of multiple strategies are investigated in [2, 15, 7].”