Distributed Data Management

This topic investigates decentralized, data-driven management of tasks in Edge Computing environments. Methods dealing with query and task allocation accross data centers and computing nodes are intorduced based on:

  1. Quality-driven task offloading mechanisms and 
  2. Time-optimized task offloading mechanisms based on the Theory of Optimal Stopping. 

Moreover, this topic covers probabilistic methods for proactive service management and service placement (vNFs) at the network edge including monitoring of the Quality-of-Service (QoS) and Quality-of-Experience (QoE).

Selected outcomes from this topic have been presented and published in: [IEEE TNSM], [AIAI 2002], [ComNet 2022], [IEEE ICTAI 2021], [IEEE FiCloud 2022], [Smart Cities], [Evolving Systems], [ACM Computing Surveys], [FGCS 2021a], [FGCS 2021b], [IM 2021], [Internet of Things 2020], [ACM TOIT 2020], [Computing 2020], [GlobeCom 2020], [IEEE ISCC 2020], [JNCA 2019], [IEEE IDCS 2019], [FGCS 2019], [IEEE INFOCOM 2018], [IEEE ICC 2018

Resilient Distributed Learning & Reusability

The topic of distributed learning paradigms deals with novel Federated Learning (FL) methods in distributed computing environments, including personalized FL and model approximation and sparsification at the Network Edge. 

Moreover, this topic includes the novel paradigm of Model Resuability, which focuses on the reuse of models in Edge Computing environments for inferential and predictive analytics services.

In addition, this topics investigates time-optimized ML model maintenance due to e.g., concept drifts, data streams, and contextual changes in Distributed ML systems. Our current research deals with the novel method distributed statistical and machine learning synopses management. Under this context, Resilient ML Systems mechanisms are explored and introduced.

Furthermore, contextual Multi-Armed Bandits (MAB) sequential learning methods are investigated to support fundamental tasks like next-query prediction, dynamic pricing, and novelty detection.

Selected outcomes from this topic have been presented and published in [IEEE TKDE 2022] [JNCA 2022], [FGCS 2022], [ACM ICTIR 2022], [WF-IoT 2022], [ICICS 2021a], [ICICS 2021b] [ACM/ACML 2021], [IEEE ISPDC 2019], [IEEE FiCloud 2027]

Large-scale Data Analytics

This topic investigates methods of pushing data science at the network edge supporting predictive and inferencial analytics. Moreover, the topic includes the novel method of query-driven analytics coping with large-scale distributed data. 

Paradigms of mining user-interest data regions and explainability of exploratory analytics are introduced including adaptive learning of aggregate analytics over Big Data systems. The fundamental contributions of this topic are in the following areas: 

  • Inferential & Predictive Analytics
  • Mining Interesting Data Regions
  • Query-driven Explanatory Analytics
  • Large-scale Regression
  • Adaptive Learning via Analytics

Selective publications of this topics can be found in the following conferences and journal articles: [ACM TKDD 2020], [JNCA 2020], [FGCS 2020], [IEEE ICDE 2020], [Applied Intelligence 2020], [IEEE BigData 2019], [JDSA 2020], [IEEE BigData 2018a], [IEEE BigData 20218b], [IEEE BigData 2018c], [IEEE EDGE 2018], [Applied intelligence 2018], [Evolving Systems 2018], [IEEE BigData 2017], [IEEE ICDCS 2017], [ACM TKDD 2017], [IEEE ICDE 2017], [Big Data Research 2015], [IEEE ICDM 2015], [IEEE Big Data 2015], [Big Data 2015], [ACM KDD 2014]

Information Processing Systems

The contextual information processing topic covers mechanims for in-network information processing in mobile Wireless Sensor Networks (WSNs), Unmanned Vehicles networks and ad-hoc networks. The topic focuses on developping and deploying Context-aware Computing Systems ranging from context-aware modelling, context reasoning/inference, to contextual information processing coming from sensors, data streams and mobile computing nodes (like vehicles, end-users, etc).

Moreover, this topic deals with the quality of capturing, deliverying and managing contextual information introducing paradigms in Delay-tolerant Networks and Mobile Computing environments.

Selected outcomes of this topic have been published in conferences and journal articles: [IFIP Networking 2021], [IM 2021], [MSWiM 2020], [ACM TOIT], [Information 2019], [Wireless Networks 2019], [WiMOB 2018], [ComCom], [IEEE TSMC], [Info Sci], [ACM TOSN], [JWIN], [IEEE Internet of Things], [Inf Sci], [ComNet], [ESWA], [ACM TOSN], [IEEE TSMC(a)], [IEEE TSMC(b)], [JPDC], [PUC], [IEEE MDM], [IEEE PIMRC], [ComNet]