MLJ 2016 - Special Issue of MLJ on Metalearning and Algorithm Selection

Papers are solicited describing recent work in the area of algorithm selection and configuration which arises in many diverse domains, such as machine learning, data mining, optimization and satisfiability solving. Metalearning leverages knowledge of past algorithm applications for selecting the best techniques for the current problem. The aim of this call is to gather submissions that discuss diverse approaches for the algorithm selection and configuration problem, with the aim of identifying the potentially best algorithm(s) for a new task, based on meta-level information and prior or current experiments. Many contemporary problems require that solutions be elaborated in the form of workflows which include different processes or operations, including configuration of algorithms or preprocessing, besides simple algorithm selection. Constructing such workflows requires extensive expertise. Contributions are welcome that address how approaches to algorithm selection and configuration could be extended to this more challenging setting. Potential authors may consider the following list of topics as guidance when preparing the submission. This list is not exhaustive, as other topics that are strongly associated with algorithm selection and metalearning may also be considered.
? Algorithm / Model selection and configuration
? Meta-learning and exploitation of meta-knowledge
? Experimentation and evaluation of learning processes
? Hyper-parameter optimization
? Planning to learn and to construct workflows
? Exploitation of ontologies of tasks and methods
? Exploitation of benchmarks and experimentation
? Representation of learning goals and states in learning
? Control and coordination of learning processes
? Meta-reasoning
? Layered learning
? Multi-task and transfer learning
? Learning to learn
Deadlines
? Paper submission: 10 February 2016
? Acceptance decision: 31 April 2016
Submission process
Authors should follow the standard procedure for submitting papers to MLJ, and use the Editorial Manager (EM) (http://MACH.edmgr.com/), identifying Article Type as Metalearning and Algorithm Selection.
Editors
? Pavel Brazdil, University of Porto, Portugal, pbrazdil-AT-inescporto.pt
? Christophe Giraud-Carrier, Brigham Young University, USA
Reviewing process and reviewers
An experienced team of reviewers with expertise in metalearning will carry out the review of all papers according to MLJ standards. All reviewing will be done in collaboration with the Journal Editorial Office (JEO).