Hierarchical sampling for active learning

By Sanjoy Dasgupta, Daniel J. Hsu. 2008

526 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Mechanical Engineering

How researchers see this paper

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

In Artificial Intelligence

In Mechanical Engineering

Citations in the corpus

The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition (2015, neutral): “In contrast to our work, the oracle is often simulated with either a simple or no noise model [4, 32, 10, 55].”
Replay in Deep Learning: Current Approaches and Missing Biological Elements (2021, neutral): “…Wang, Zhang, Li, Zhang, & Lin, 2016; Settles, 2009; Yoo & So Kweon, 2019; Wei, Iyer, & Bilmes, 2015), with the most common selection methods using uncertainty sampling (Lewis & Gale, 1994; Culotta & McCallum, 2005; Scheffer, Decomain, & Wrobel, 2001; Dagan & Engelson, 1995; Dasgupta & Hsu, 2008).”
Semisupervised Classification of Remote Sensing Images With Active Queries (2012, methodology): “select the most informative pixels to be labeled [12], [13].”
A Graphbased Framework for Multi-Task Multi-View Learning (2011, methodology): “In Section 4, we introduced the objective function for GraM 2 , i.e., Equation (2).”
Importance weighted active learning (2008, methodology): “In such cases, they can exhibit a lack of statistical consistency: even with an inﬁnite labeling budget, they might not converge to an optimal predictor (see Dasgupta and Hsu (2008) for a discussion).”
Active Learning Literature Survey (2009, background): “However, some recent theoretical work has begun to address these issues, coupled with promising empirical results (Dasgupta and Hsu, 2008; Beygelzimer et al., 2009).”
Semi-automated screening of biomedical citations for systematic reviews (2010, background): “Finally, exploiting latent hierarchical structure in the UMLS ontologies might also be helpful, particularly because active learning techniques for hierarchical data have recently been developed [38].”
Hierarchical Subquery Evaluation for Active Learning on a Graph (2014, methodology): “The hierarchical clustering, in [6], is used to deﬁne bounds on sampling statistics.”
Cost-Effective Incremental Deep Model: Matching Model Capacity With the Least Sampling (2023, unknown): “Classical approaches include random sampling, uncertainty sampling [39], [40], query-by-commit-tee [41], hierarchical sampling [42] and so on.”
Generative Adversarial Active Learning (2017, methodology): “Some pool-based algorithms make selections by using clustering techniques or maximizing a diversity measure, e.g. (Brinker; Xu et al., 2007; Dasgupta and Hsu, 2008; Nguyen and Smeulders; Yang et al., 2014; Hoi et al., 2009).”
Two faces of active learning (2011, background): “A rudimentary label complexity result for this model is proved in [11]: if the provided hierarchical clustering contains a pruning P whose clusters are ǫ -pure in their labels, then the learner will ﬁnd a labeling that is O ( ǫ )-pure with O ( P | d ( P ) /ǫ ) labels, where d ( P ) is the maximum…”
Group-Based Active Learning of Classification Models (2017, methodology): “As mentioned in introduction section, however, instance-based active learning framework could suffer from the sampling bias problem (Dasgupta and Hsu 2008) and it is unfriendly to high-dimensional queries.”
A brief introduction to weakly supervised learning (2018, background): “Approaches based on representativeness generally aim to exploit the cluster structure of unlabeled data, usually by a clustering method [23, 57].”
A Lazy Man's Approach to Benchmarking: Semisupervised Classifier Evaluation and Recalibration (2013, background): “For example, is it possible to do unbiased active querying, so that the oracle is asked to label the most informative examples? One possibility in this direction would be to employ importance weighted active sampling techniques [3, 6], so similar in spirit to [19, 18] but for performance curves.”