The Supervised Learning No-Free-Lunch Theorems

By David Hilton Wolpert. 2002

531 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/689c74bd55e0be2b0866b9a86ec53363ce82aed1

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

Citations in the corpus

An Algorithm Based on Deep Learning for Predicting In‐Hospital Cardiac Arrest (2018, result): “Wolpert explains the “No Free Lunch” theorem: If optimized in one situation, an algorithm cannot produce good results in other situations.(43) To overcome this issue, we experimented in 2 ways.”
Disease Diagnosis in Smart Healthcare: Innovation, Technologies and Applications (2017, background): “There is a “No free lunch” theorem in machine learning which says that no unique algorithm works best for every application [176].”
Advanced nonparametric tests for multiple comparisons in the design of experiments in computational intelligence and data mining: Experimental analysis of power (2010, background): “It is not possible to find one algorithm that is the best in behavior for all problems, as the ''no free lunch " theorem suggests [50] [51].”
Multicenter Comparison of Machine Learning Methods and Conventional Regression for Predicting Clinical Deterioration on the Wards (2016, background): “Thus, as per Wolpert’s “No Free Lunch Theorem,” no one technique will be most accurate across every scenario, and so comparisons of techniques in different research areas and datasets may yield different results (24).”
Ten Simple Rules for Reducing Overoptimistic Reporting in Methodological Computational Research (2015, background): “Think of the widely acknowledged “no-free-lunch theorem” [11].”
A survey of multiple classifier systems as hybrid systems (2014, background): “Wolpert stated that each classifier has its specific competence domain [3], where they overcome other competing algorithms, thus it is not possible to design a single classifier which outperforms another ones for each classification tasks.”
Multi-target regression via input space expansion: treating targets as inputs (2012, methodology): “and add a bias column as suggested inZhou et al. (2011b). Finally,Dirty is setup as suggested in Jalali et al.”
Review of Bridge Structural Health Monitoring Aided by Big Data and Artificial Intelligence: From Condition Assessment to Damage Detection (2020, unknown): “Though many classification algorithms are available, it is still impossible to find an optimal classifier with excellent performance in each aspect; such a phenomenon in supervised learning is also known as the no free lunch theorem (Wolpert 2002).”
Cost-sensitive decision tree ensembles for effective imbalanced classification (2014, unknown): “However, it is also well known that according to the no free lunch theory there is no universal classifier that performs best for all decision problems [41].”
Reservoir computing approaches to recurrent neural network training (2009, background): “It is worth mentioning at this point that the general “no free lunch” principle in supervised machine learning [62] states that there can exist no bias of a model which would universally improve the accuracy of the model for all possible problems.”
A Plea for Neutral Comparison Studies in Computational Sciences (2012, methodology): “so called ‘‘no-free lunch theorem’’ [4] in mind, we do not expect papers claiming that one of the methods, for instance the new”
Ensemble learning for data stream analysis: A survey (2017, methodology): “The main motivation for using classifier ensembles is the no free lunch theorem formulated by Wolpert [186].”
Incremental on-line learning: A review and comparison of state of the art algorithms (2018, methodology): “An algorithm has to be chosen according to the preconditions of a given task since there cannot exist a method which optimally performs in every scenario [5].”
No Free Lunch Theorem: A Review (2019, background): “Revisiting his initial work on supervised learning Wolpert in his work entitled: “The supervised learning no-free-lunch theorems” [59] analyzes the main issues underlying his theory on NFL.”
Algorithm Selection for Combinatorial Search Problems: A Survey (2012, background): “ solve has the potential to improve performance. The theorems would apply to Algorithm Selection systems themselves as well though (in particular the version for supervised learning are relevant, see Wolpert, 2001). This means that although performance improvements can be achieved by selecting the r”