A5070508409- Service-Oriented Architecture and Web Services
- Advanced Software Engineering Methodologies
- Business Process Modeling and Analysis
- Time Series Analysis and Forecasting
- Advanced Database Systems and Queries
- Stock Market Forecasting Methods
- Advanced Text Analysis Techniques
- Software System Performance and Reliability
- Data Quality and Management
- Bayesian Modeling and Causal Inference
- Anomaly Detection Techniques and Applications
- Neural Networks and Applications
- Network Security and Intrusion Detection
- Multimodal Machine Learning Applications
- Statistical Methods and Inference
- Logic, programming, and type systems
- Semantic Web and Ontologies
- Data Mining Algorithms and Applications
- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Statistical Methods and Bayesian Inference
- Data Stream Mining Techniques
Bundesministerium für Klimaschutz, Umwelt, Energie, Mobilität, Innovation und Technologie
2018-2021
Laboratoire d'Informatique de Grenoble
2007-2010
Université Grenoble Alpes
2007-2010
Interactions between two applications encapsulated into Web services consist in series of message exchanges that must conform to service interfaces. The study reported this text aims at dealing with the issues arise when interactions (a client and a provider) fail because their interfaces are incompatible. This may happen provider has evolved its interface been modified. It also decided change for another which addresses same needs but offers different interface. contribution proposed...
Information technology (IT) systems are vital for modern businesses, handling data storage, communication, and process automation. Monitoring these is crucial their proper functioning efficiency, as it allows collecting extensive observational time series analysis. The interest in causal discovery growing IT monitoring knowing relations between different components of the system helps reducing downtime, enhancing performance identifying root causes anomalies incidents. It also proactive...
This paper introduces a new structural causal model tailored for representing threshold-based IT systems and presents algorithm designed to rapidly detect root causes of anomalies in such systems. When are not causally related, the method is proven be correct; while an extension proposed based on intervention agent relax this assumption. Our its agent-based leverage discovery from offline data engage subgraph traversal when encountering online data. extensive experiments demonstrate superior...
When two Web services work together, they exchange messages in a predefined interface process. Two processes should be compatible when can properly. Our idea to fix incompatibility problem service is change an incompatible process so that the new simulate We consider not only control flow but also data modeling into FSMs. present technique detects incompatibility, provides resolution strategies generate
In this paper, we describe a method to use graph-based language modeling approach for image retrieval and categorization. We first mapped regions induced concepts then spatial relationships between these build graph representation of images. Our allows deal with different scenarii, where isolated images or groups are used training testing. The results obtained on an categorization problem comprising 3849 from 101 landmarks Singapore show that (a) the procedure automatically induce is...
Interactions between Web services are based on interfaces which describe both structural and behavioural perspectives. It can happen that the interface provided by a service does no longer match (for instance, because of an evolution) required its partners. In this situation, until fixed, interactions cannot succeed. To address issue, focusing part interfaces, we propose approach mediator, automatically generated, aims to seamlessly resolve incompatibilities during interactions.
Monitoring thousands of machines and services in a datacenter produces lot time series points, giving general idea the health cluster. However, there is lack tools to further exploit this data, for instance prediction purposes. We propose apply linear regression algorithms predict future behavior monitored systems anticipate downtimes, system administrators information they need ahead problems arising. This problem quite challenging when dealing with high number monitoring metrics, given our...
Interactions between Web services are based on interfaces which describe both structural and behavioural perspectives. It can happen that the interface provided by a service does no longer match (for instance, because of an evolution) required its partners. In this situation, until fixed, interactions cannot succeed. To address issue, focusing part interfaces, we propose approach mediator, automatically generated, aims to seamlessly resolve incompatibilities during interactions.