• Corpus ID: 17130634

Metalearning - A Tutorial

  title={Metalearning - A Tutorial},
  author={Christophe G. Giraud-Carrier},

Figures and Tables from this paper

An intelligent decision support system for machine learning algorithms recommendation

This work proposes an application that might help deciding on which machine learning methods a user needs for a specified problem, and presents a prototype of such a system, and the implementation details of the two recommender algorithms.

A Study of Algorithm Selection in Data Mining using Meta-Learning

This study has analyzed the working of a metalearning system in connection with the classical algorithm selection problem and identifies few potential challenges in relation to algorithm selection in data mining using meta-learning.

Meta-Learning and the Full Model Selection Problem

This thesis presents research results, including new methods, algorithms, and theoretical and empirical analysis from two directions, which try to propose systematic and efficient solutions to the questions above, using different resource requirements, namely, the meta-learning-based algorithm/parameter ranking approach and theMeta-heuristic search-based full-model selection approach.

Electric Load Forecasting using Multivariate Meta- learning

The modern economies depend on electricity and its production and consumption (load) have to be in equilibrium at all times since storing electricity produces electricity and consumes electricity.

Pairwise meta-rules for better meta-learning-based algorithm ranking

This paper presents a novel meta-feature generation method based on rules that compare the performance of individual base learners in a one-against-one manner, and introduces a new meta-learner called Approximate Ranking Tree Forests (ART Forests) that performs very competitively when compared with several state-of-the-art meta-learners.

Application of meta-learning in cyberspace security: A survey

Meta‐learning and the new challenges of machine learning

A framework for dealing with challenges of automatic algorithm selection and continuous algorithm updates that do not require the retraining of the whole algorithm to adapt to new data is proposed, based on meta‐learning.

Critical infrastructure protection based on memory-augmented meta-learning framework

  • Xie Bing
  • Computer Science
    Neural Computing and Applications
  • 2020
This research proposes a novel meta-learning architecture that considers the neural turing machines as the approach upon which the model is founded and facilitates the rapid integration of new information without the need for retraining.

OpenML-Connect: A C++ connection library for OpenML

A C++ library that contains a connector to easily connect to OpenML’s API and access the data it provides and it is found that the trained model performed better than two alternative baseline methods that were used in this study.


Este artigo tem como objetivo estudar as caracteristicas de certas bases de dados para a selecao de algoritmos de modelagem atraves de medidas de complexidade propostas por Lorena et al. (2018). Com



A Treatise of Human Nature (review)

ideas at; the list of predecessors' views on the causal link at; the listing of sources for skeptical attacks on the reliability of the senses at; the amusing reminder of

Odd bites into bananas don''t make you blind: learning about simplicity and attribute addition

A meta-learning technique that uses unpruned decision trees and information from decision tree construction to describe the learning tasks is described and results suggest that induction on the connection between problems and simplicity and representation biases improves learning performance.

A procedure for ranking efficient units in data envelopment analysis

Data Envelopment Analysis DEA evaluates the relative efficiency of decision-making units DMUs but does not allow for a ranking of the efficient units themselves. A modified version of DEA based upon

Shift of bias for inductive concept learning

It is shown that search for appropriate bias is itself a major part of the learning task, and that mechanical procedures for conducting a well-directed search for an appropriate bias can be created.

A Proposal for Meta-Learning Through a Multi-Agent System

The MAS-based information system the authors use for mining and meta-learning is explained, and how the scalability problem is solved in order to support a community of many software agents is explained.

Improving Progressive Sampling via Meta-learning on Learning Curves

The paper presents a method for predicting the stopping point using a meta-learning approach that requires just four iterations of the progressive sampling and can lead to significant savings of time without significant losses of accuracy.

Characterization of Classification Algorithms

This paper is concerned with the problem of characterization of classification algorithms and involves generation of different kinds of models, which include regression and rule models, piecewise linear models (model trees) and instance based models.

Sampling-Based Relative Landmarks: Systematically Test-Driving Algorithms Before Choosing

Sampling-based landmarks (SL), a systematization of this approach, building on earlier work on landmarking and sampling, are described, that address the inability of earlier landmarks to assess relative performance of algorithms.