A Scalable Test Suite for Continuous Dynamic Multiobjective Optimization

Jiang, Shouyong; Kaiser, Marcus; Yang, Shengxiang; Kollias, Stefanos; Krasnogor, Natalio

doi:10.1109/tcyb.2019.2896021

Cited by 38 publications

(52 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We are currently extending the developed approach for unsupervised domain adaptation, using Bayesian formulations to provide the networks with selftraining capabilities [28], while formulating it as a multi-objective optimization problem [11]. We are also investigating the use of capsules in the developed algorithms [6,29] so as to include feedback and model structuring in the generated models and representations.…”

Section: Discussionmentioning

confidence: 99%

Transparent Adaptation in Deep Medical Image Diagnosis

Kollias

Vlaxos

Seferis

et al. 2021

Trustworthy AI - Integrating Learning, Optimization and Reasoning

View full text Add to dashboard Cite

The paper presents a novel deep learning approach, which extracts latent information from trained Deep Neural Networks (DNNs) and derives concise representations that are analyzed in an effective, transparent way for prediction in medical imaging. A novel methodology is presented, in which deep neural architectures that have been trained to provide highly accurate predictions over existing datasets are adapted, in a consistent way, to make predictions over different contexts and datasets. Unified prediction is then achieved over the original and the new datasets. Successful application is illustrated through a large experimental study for prediction of Parkinson's disease from MRI and DaTScans, as well as for prediction of COVID-19 from CT scans and X-rays.

show abstract

Section: Discussionmentioning

confidence: 99%

Transparent Adaptation in Deep Medical Image Diagnosis

Kollias

Vlaxos

Seferis

et al. 2021

Trustworthy AI - Integrating Learning, Optimization and Reasoning

View full text Add to dashboard Cite

show abstract

“…For this case the constraint in Eq. (22) was defined as ξ(x) = 2. g(x). The dissimilarity of the functions can also be controlled.…”

Section: Equality and Inequality Constraintsmentioning

confidence: 99%

“…In addition to the possibilities listed above, the constraints presented in Eqs. (19) to (22) can be added, as well as the dissimilarity of objectives as the presented in Eq. (8).…”

Section: Proposed Generator Of Benchmark Problemsmentioning

confidence: 99%

Scalable and customizable benchmark problems for many-objective optimization

Meneghini

Alves

Gaspar‐Cunha

et al. 2020

Applied Soft Computing

View full text Add to dashboard Cite

Solving many-objective problems (MaOPs) is still a significant challenge in the multi-objective optimization (MOO) field. One way to measure algorithm performance is through the use of benchmark functions (also called test functions or test suites), which are artificial problems with a well-defined mathematical formulation, known solutions and a variety of features and difficulties. In this paper we propose a parameterized generator of scalable and customizable benchmark problems for MaOPs. It is able to generate problems that reproduce features present in other benchmarks and also problems with some new features. We propose here the concept of generative benchmarking, in which one can generate an infinite number of MOO problems, by varying parameters that control specific features that the problem should have: scalability in the number of variables and objectives, bias, deceptiveness, multimodality, robust and non-robust solutions, shape of the Pareto front, and constraints. The proposed Generalized Position-Distance (GPD) tunable benchmark generator uses the position-distance paradigm, a basic approach to building test functions, used in other benchmarks such as Deb, Thiele, Laumanns and Zitzler (DTLZ), Walking Fish Group (WFG) and others. It includes scalable problems in any number of variables and objectives and it presents Pareto fronts with different characteristics. The resulting functions are easy to understand and visualize, easy to implement, fast to compute and their Pareto optimal solutions are known.

show abstract

“…Considering the varying forms of the POS and POF, DMOPs generally are classified into four different types [1,17]:…”

Section: Definition 2 Pareto Optimal Set (Pos)mentioning

confidence: 99%