Corpus ID: 29941798

Robust learning in safety related domains: machine learning methods for solving safety related application problems

  title={Robust learning in safety related domains: machine learning methods for solving safety related application problems},
  author={S. Nusser},
Today, machine learning methods are successfully deployed in a wide range of applications. A multitude of different learning algorithms has been developed in order to solve classification and regression problems. These common machine learning approaches are regarded with suspicion by domain experts in safety-related application fields because it is often infeasible to sufficiently interpret and validate the learned solutions. Especially for safety-related applications, it is imperative to… Expand
Gas turbine engine prognostics using Bayesian hierarchical models: A variational approach
Abstract Prognostics is an emerging requirement of modern health monitoring that aims to increase the fidelity of failure-time predictions by the appropriate use of sensory and reliabilityExpand
Bayesian Approaches for Complex System Prognostics
An integration concept is proposed, combining the Bayesian data modelling technique with an information theoretic change-point detection algorithm to solve a wide class of prognostic problems, such as information arising from irregular events occurring during the life-cycle of an asset. Expand
Exploiting Robust Multivariate Statistics and Data Driven Techniques for Prognosis and Health Management
This thesis explores state of the art robust multivariate statistical methods and data driven techniques to holistically perform prognostics and health management (PHM). This provides a means toExpand
Learning and decision-making in artificial animals
The animat framework enables a uniform and gradual approach to AGI, by successively taking on more challenging problems in the form of broader and more complex classes of environments. Expand
Evolutionary Fuzzy Rules for Ordinal Binary Classification with Monotonicity Constraints
  • C. Moewes, R. Kruse
  • Computer Science
  • Soft Computing: State of the Art Theory and Novel Applications
  • 2013
An approach to learn fuzzy binary decision rules from ordinal temporal data where the task is to classify every instance at each point in time by using the Variable Consistency Dominance-based Rough Set Approach (VC-DRSA) to exploit preference information about the problem. Expand
ERTMS Challenges for a Safe and Interoperable European Railway System
Without an overarching process, cross-discipline understanding and improved ascribing meaning to data, implementing ERTMS does not mean the railway system will become safer, and effects of deregulation, dynamic specifications, interoperability and time drain make that allowing an interoperable railway system by implementing ERtMS appears not to be self-explanatory. Expand
PESARO 2017 Proceedings
Falling of the elderly has become an important issue in today's aging society. As a result, active protecting devices are being developed to protect the fallers’ body from severe injuries. The keyExpand
Herausforderungen in der Absicherung von Fahrerassistenzsystemen bei der Benutzung maschinell gelernter und lernender Algorithmen
Zusammenfassung: Der Einsatz maschinell gelernter und lernender Algorithmen verstärkt sich im Bereich der Fahrerassistenzsysteme zunehmend basierend auf den hieraus resultierenden Vorteilen wie derExpand


Multi-Class Extension of Verifiable Ensemble Models for Safety-Related Applications
This contribution discusses two different strategies for extending a verifiable ensemble approach for binary classification tasks to also solve multi-class problems by introducing a hierarchy of misclassification costs. Expand
Learning Verifiable Ensembles for Classification Problems with High Safety Requirements
Machine learning methods are successfully applied in a wide range of applications – for instance – object recognition in computer vision, search engines, or stock market analysis. But in the field ofExpand
Interpretable ensembles of local models for safety-related applications
A feature construction method based on a multi-layer percep- tron architecture is proposed to overcome limitations of the local modeling strategy, while keeping the global model interpretable. Expand
Developing artificial neural networks for safety critical systems
The safety criteria which are safety requirements for the behaviour of neural networks are outlined and can contribute to justifying the safety of ANN functional properties and a safety lifecycle for artificial neural networks is presented. Expand
Using Fuzzy Self-Organising Maps for Safety Critical Systems
A type of constrained Artificial Neural Network that enables analytical certification arguments whilst retaining valuable performance characteristics whilst retaining value performance characteristics is defined. Expand
Verification of a trained neural network accuracy
  • R. Zakrzewski
  • Computer Science
  • IJCNN'01. International Joint Conference on Neural Networks. Proceedings (Cat. No.01CH37222)
  • 2001
A deterministic method for verification of a neural net trained to approximate another, difficult to implement mapping, and allows us to verify accuracy of nets that replace large multidimensional look-up tables. Expand
Verifiable Ensembles of Low-Dimensional Submodels for Multi-class Problems with Imbalanced Misclassification Costs
This chapter proposes a solution of this problem that avoids such inconsistencies by introducing a hierarchy of misclassification costs and shows that by following such a hierarchy it becomes feasible to extend thebinary ensemble, to maintain the desirable properties of the binary ensemble, and to achieve a good predictive performance. Expand
Robust Classification for Imprecise Environments
It is shown that it is possible to build a hybrid classifier that will perform at least as well as the best available classifier for any target conditions, and in some cases, the performance of the hybrid actually can surpass that of the best known classifier. Expand
Obtaining interpretable fuzzy classification rules from medical data
Extensions to the learning algorithms of neuro-fuzzy classification (NEFCLASS), a neuro- fuzzy approach for data analysis that has been presented before, are discussed and interactive strategies for pruning rules and variables from a trained classifier to enhance its readability are presented. Expand
Machine learning
Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Expand