Hoeffding Races: Accelerating Model Selection Search for Classification and Function Approximation

By Oded Maron, Andrew W. Moore. 1993

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

Multi-Fidelity Automatic Hyper-Parameter Tuning via Transfer Series Expansion (2019, background): “Model conﬁguration is often considered as model selection (Brazdil, Soares, and Da Costa 2003; Chapelle, Vap-nik, and Bengio 2002; Maron and Moore 1994; Zhao and Yu 2006) and hyper-parameter tunning (Bergstra and Ben-gio 2012; Snoek, Larochelle, and Adams 2012).”
Benchmark and Survey of Automated Machine Learning Frameworks (2019, background): “A quite simple approximation is to abort the fitting after the first fold if the performance is significantly worse than the current incumbent (Maron & Moore, 1993; Hutter et al., 2011).”
ParamILS: An Automatic Algorithm Configuration Framework (2014, methodology): “Another approach is based on adaptations of racing algorithms in machine learning (Maron & Moore, 1994) to the algorithm conﬁguration problem.”
Preference-based Evolutionary Direct Policy Search (2013, methodology): “The racing framework is an uncertainty handling scheme introduced in [9], [10].”
Taking Human out of Learning Applications: A Survey on Automated Machine Learning (2018, background): “Such configurations can usually be easily identified at the early stage of model training, with their performance monitored on the validation set [65], [105], [129], [130].”
A Survey of Preference-Based Online Learning with Bandit Algorithms (2014, background): “In principle, these learning problems are very similar to the value-based racing task [38, 39], where the goal is to select the k arms with the highest means.”
Locally Weighted Learning for Control (1997, background): “Cheap cross validation makes search for model parameters routine, and we have explored procedures that take advantage of this (Atkeson et al. 1997; Maron and Moore 1994; Moore et al. 1992; Moore and Lee 1994).”
Preference-based Online Learning with Dueling Bandits: A Survey (2018, background): “In principle, these learning problems are very similar to the value-based racing task (Maron and W., 1994; Maron and Moore, 1997), where the goal is to select the k arms with the highest means.”
Best Arm Identification: A Unified Approach to Fixed Budget and Fixed Confidence (2012, background): “Maron & Moore [12] considered a slightly different setting where besides a fixed confidence also the maximum number of rounds is fixed.”
On expressiveness and uncertainty awareness in rule-based classification for data streams (2017, methodology): “The use of Hoeﬀding Inequality in Hoeﬀding Rules was inspired by [44, 45, 10].”
Automated Algorithm Selection: Survey and Perspectives (2018, background): “These run candidate algorithms on an increasing number of instances, eliminating those from consideration whose performance is signiﬁcantly below that of others, based on a statistical test (see, e.g., Maron and Moore, 1994; Birattari et al., 2002).”
Speeding Up Automatic Hyperparameter Optimization of Deep Neural Networks by Extrapolation of Learning Curves (2015, methodology): “Other approaches proposed over the years include racing algorithms [Maron and Moore, 1994] and gradient search [Bengio , 2000].”
Beyond Manual Tuning of Hyperparameters (2015, methodology): “Other approaches proposed over the years include gradient search [4], population-based search [26] and racing algorithms [44].”
On correlation and budget constraints in model-based bandit optimization with application to automatic machine learning (2014, background): “, 2013], and go back to old proposals for classifier selection [Maron and Moore, 1994].”
Machine Learning for Sequential Data: A Review (2002, methodology): “Several methods have been developed that identify such interactions including RELIEFF [14], Markov blankets [13], and feature racing [17].”