Analysis of Perceptron-Based Active Learning

Dasgupta, Sanjoy; Kalai, Adam Tauman; Monteleoni, Claire

doi:10.1007/11503415_17

Cited by 123 publications

(161 citation statements)

References 9 publications

Supporting

Mentioning

155

Contrasting

Unclassified

Order By: Relevance

“…Let C denote the class that contains the empty set and the class of disjoint consecutive intervals defined in (14). For sake of simplicity assume that n/s is an integer.…”

Section: S-intervalsmentioning

confidence: 99%

“…In statistics, these are often referred to as sequential design of experiments [29,10,18,17,20,5,24,25]. In machine learning and computer science, these methodologies are often known as active learning, and are of great importance in applied settings [27,11,12,13,14,30,7,4,23,21]. The effect that structural information regarding the unknown parameter of interest has on statistical inference is an important question, and was studied in e.g.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Adaptive Sensing for Estimation of Structured Sparse Signals

Castro

Tánczos

2015

IEEE Trans. Inform. Theory

View full text Add to dashboard Cite

In many practical settings one can sequentially and adaptively guide the collection of future data, based on information extracted from data collected previously. These sequential data collection procedures are known by different names, such as sequential experimental design, active learning or adaptive sensing/sampling. The intricate relation between data analysis and acquisition in adaptive sensing paradigms can be extremely powerful, and often allows for reliable signal estimation and detection in situations where non-adaptive sensing would fail dramatically. In this work we investigate the problem of estimating the support of a structured sparse signal from coordinate-wise observations under the adaptive sensing paradigm. We present a general procedure for support set estimation that is optimal in a variety of cases and shows that through the use of adaptive sensing one can: (i) mitigate the effect of observation noise when compared to non-adaptive sensing and, (ii) capitalize on structural information to a much larger extent than possible with non-adaptive sensing. In addition to a general procedure to perform adaptive sensing in structured settings we present both performance upper bounds, and corresponding lower bounds for both sensing paradigms.

show abstract

“…Let C denote the class that contains the empty set and the class of disjoint consecutive intervals defined in (14). For sake of simplicity assume that n/s is an integer.…”

Section: S-intervalsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Adaptive Sensing for Estimation of Structured Sparse Signals

Castro

Tánczos

2015

IEEE Trans. Inform. Theory

View full text Add to dashboard Cite

show abstract

“…Active learning in an online setting has been discussed in [5] and [3]. The work of [5] assumes input data points uniformly distributed over the surface of an unit sphere.…”

Section: Related Workmentioning

confidence: 99%

“…The work of [5] assumes input data points uniformly distributed over the surface of an unit sphere. However, we cannot make such distributional assumptions for domain adaptation.…”

Section: Related Workmentioning

confidence: 99%

Active Supervised Domain Adaptation

Saha

Rai

Daumé

et al. 2011

Machine Learning and Knowledge Discovery in Databases

View full text Add to dashboard Cite

Abstract. In this paper, we harness the synergy between two important learning paradigms, namely, active learning and domain adaptation. We show how active learning in a target domain can leverage information from a different but related source domain. Our proposed framework, Active Learning Domain Adapted (Alda), uses source domain knowledge to transfer information that facilitates active learning in the target domain. We propose two variants of Alda: a batch B-Alda and an online O-Alda. Empirical comparisons with numerous baselines on real-world datasets establish the efficacy of the proposed methods.

show abstract

“…Our model has obvious connections to active learning (Seung et al 1992;Kulkarni et al 1993;Freund et al 1997;Liere and Tadepalli 1997;Fine et al 2002;Dasgupta et al 2005;Dasgupta 2004Dasgupta , 2005. In some sense the output identification task is much simpler than the usual learning task, since we do not need to find an accurate hypothesis, only to target one example for each output value.…”

mentioning

confidence: 93%

Active sampling for multiple output identification

Fine

Mansour

2007

Mach Learn

View full text Add to dashboard Cite

We study functions with multiple output values, and use active sampling to identify an example for each of the possible output values. Our results for this setting include: (1) Efficient active sampling algorithms for simple geometric concepts, such as intervals on a line and axis parallel boxes. (2) A characterization for the case of binary output value in a transductive setting. (3) An analysis of active sampling with uniform distribution in the plane. (4) An efficient algorithm for the Boolean hypercube when each output value is a monomial.

show abstract

Analysis of Perceptron-Based Active Learning

Cited by 123 publications

References 9 publications

Adaptive Sensing for Estimation of Structured Sparse Signals

Adaptive Sensing for Estimation of Structured Sparse Signals

Active Supervised Domain Adaptation

Active sampling for multiple output identification

Contact Info

Product

Resources

About