Queries and concept learning

By Dana Angluin. 1988

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

Semi-supervised Knowledge Transfer for Deep Learning from Private Training Data (2016, background): “Active learning is a class of techniques that aims to identify and prioritize points in the student’s training set that have a high potential to contribute to learning (Angluin, 1988; Baum, 1991).”
Stealing Machine Learning Models via Prediction APIs (2016, background): “This approach would place model extraction back in the membership query setting of prior work in learning theory [3, 8, 36, 53].”
Syntax-guided synthesis (2013, methodology): “We refer the reader to a paper by Angluin [23] for an overview of various models for query-based active learning.”
Recent Developments in Algorithmic Teaching (2009, background): “This is even true for the query learning model introduced by Angluin [1, 2], since the teacher or oracle, though answering truthfully, is assumed to behave adversarially whenever possible.”
Active Learning and Intrinsically Motivated Exploration in Robots : Advances and Challenges (2010, background): “The field of statistical learning theory studied approaches such as optimal experimental design and active learning [1]–[6], [19], [39], and the field of developmental and cognitive robotics studied approaches based on intrinsic motivation [7], [9], [23], [24].”
On the Power of Membership Queries in Agnostic Learning (2009, methodology): “For example monotone DNF formulas, ﬁnite automata and decision trees are only known to be learnable with membership queries (Valiant, 1984; Angluin, 1988; Bshouty, 1995).”
The strategic student approach for life-long exploration and learning (2012, background): “Recent development from machine learning, mainly from active learning and multi-armed bandits, started to contribute to a formal view on the complexity of learning agents that choose their own samples: optimal experimental design and active learning [8], [15], [26]–[28], n-armed bandits [29] and the”
PRADA: Protecting Against DNN Model Stealing Attacks (2018, background): “Model extraction is conceptually similar to concept learning [3], [5] in which the goal is to learn a model for a concept using membership queries.”
Learning under p-Tampering Attacks (2018, background): “, in the membership query model of [1].”
Algorithmic and Human Teaching of Sequential Decision Tasks (2012, background): “A closely related area for the work presented in this paper is Active Learning (AL) (Angluin 1988; Settles 2010).”
Learning in the presence of concept drift and hidden contexts (2004, background): “…work presented here relates to incremental or on-line concept learning, whichhas recently received considerable attention among theoreticians (e.g., Angluin,1988; Maass, 1991; Helmbold, Littlestone, & Long, 1992) as well as practitioners(e.g., Schlimmer & Granger, 1986; Langley, Gennari, & Iba,…”
Quantum machine learning: a classical perspective (2017, background): “However, if one considers a different learning model, such as the exact learning framework developed by Angluin [36], it is possible to prove that quantum learners can be polynomially more efficient than classical learners in terms of the number of queries to the data oracle [32,37].”
From Theories to Queries (2011, background): “While Angluin (1988) outlines several potential query types in a theoretical analysis of active learning, only the commonly-used membership query has been deemed appropriate for most real-world applications.”
Learning in the Presence of Concept Drift and Hidden Contexts (1996, background): “The work presented here relates to incremental or on-line concept learning, which has recently received considerable attention among theoreticians (e.g., Angluin, 1988; Maass, 1991; Helmbold, Littlestone, & Long, 1992) as well as practitioners (e.g., Schlimmer & Granger, 1986; Langley, Gennari, & ”
Teaching Randomized Learners (2006, background): “Angluin and Kriķis’ [3, 4] model prevents collusion by giving incompatible hypothesis spaces to teacher and learner.”
Can machine learning be secure? (2006, background): “In many scenarios, it has been shown that learning is possible even in the worst case [2].”