This page contains information related to my work throughout since 2004. In particular, I try to provide details regarding my involvement in various R&D projects and efforts. Each project is briefly presented and then I try to swiftly present my involvement through videos, papers and presentations where possible. This section is divided into two parts, namely Industry and EU R&D projects
Industry
The last couple of years I am working as a Senior Data Scientist in Intracom Telecom S.A. During the same period I have collaborated in parallel as an External Advisor with promising start-ups as well as my former employer, the University of Athens. Here, I will attempt to provide a short overview of the results of these collaborations.
For two years I participated in a research effort performed in the context of a private research contract between Huawei Technologies Co., Ltd and the University of Athens. My work focused on the design, implementation and experimental assessment of a knowledge discovery framework for future 5G mobile communication systems. In particular, I designed a recommendation engine facilitating network selection based on user habits and network conditions. The engine facilitated optimal network selection based on user historic information as well as the extraction of a rules set catering for network resource usage optimization. Finally, a feedback mechanism focusing on the periodic evaluation of the rules set and the derived user classes was also designed.
While collaborating with Intelen Inc I focused on the design, implementation and experimental assessment of novel methods for knowledge discovery from energy consumption data. In particular, my work evolved around three axes; specifically:
Data preprocessing by means of dimensionality reduction in order to manage inherent problems associated with observations that are described by a large number of features.
Unsupervised methods for user and building profiling in order to facilitate the reduction of energy consumption by end-use customers.
Optimization of the context monitoring process in order to dynamically and autonomously adapt the sampling rate of the monitoring procedure user habits.
More details can be found in the following two articles:
In Trebbble I delivered a series of seminar courses on the field of Data Mining, Machine Learning and its application on the mobile ecosystem. The course was divided into three parts, namely introductory, predictive modelling and applied experimentation. Particular focus was given on the management of voluminous, high dimensional datasets. Text mining was used as a motivating example showcasing the mapping of input data to high dimensional vector spaces and the subsequent application of feature selection and feature extraction methods in order to minimize noise in the dataset and unveil latent information patterns. Finally, trainees applied the knowledge they gained on simple realistic scenarios using open-source software.
R&D Projects
From 2004 until 2013 I worked as a Research Associate in the SCAN research group of the University of Athens in the context of several, highly impacting, European research projects (FP6 and FP7) undertaking technical as well as managerial duties. In the following I provide details regarding my work in their context. You can use the following links in order to navigate:
Live Video-to-Video Supporting Interactive City Infrastructure
LiveCity (Live Video-to-Video Supporting Interactive City Infrastructure) was a CIP-Pilot Action of the European Commission. LiveCity addresses a number of communities where the citizen of a city have specific challenges which can derive benefits from the use of live interactive high definition video-to-video. These communities include emergency ambulances; hospitals, doctors, museum curators, city administrations and schools. Today live interactive high definition video-to-video is becoming possible with a range of devices and applications. These are well supported in private networks owned by enterprises where quality of service is available. For a mass market in the cities quality of services is missing in the public internet. LiveCity implements a range of pilots for city communities across the public internet. We build a wireline and 4G wireless network of 4 cities and include a right of way without interference from unwanted traffic so that each user in can experience live interactive high definition video-to-video. UoA is involved in the implementation of a telemonitoring use case with the ofthalmological clinic of the Attikon hospital. Also, UoA has the management and technical overview of the WP2 which implements the LiveCity concepts in the health sector and in particular in emergency situations (ambulances and ER units).
A video showcasing the LiveCity concepts can be found here. During the FIA 2014 conference LiveCity was awarded the Best Demonstration Award (in Greek)
Main duties and responsibilities
WP2 Management
Technical Overview of the implementation of the Telemonitoring Use Case
SmartAgri Food
The SmartAgriFood project is part of the Future Internet Public-Private Partnership (FI-PPP) program and addresses farming, agri-logistics and food awareness as a use case for this. The intelligence, efficiency, sustainability and performance of the agri-food sector can be radically enhanced by using information & decision support systems that are tightly integrated with advanced internet-based networks & services. Concurrently, the sector provides use cases for Future Internet des gn from physical layer all the way up to the service layer.
Main duties and responsibilities
Technical overview of the Greenhouse pilot implementation
Design, implementation and integration of excerpts of the Greenhouse Pilot
COoperative aNd Self growing Energy awaRe Networks
CONSERN (COoperative aNd Self growing Energy awaRe Networks) is a Specifc Targeted Research Project (STREP) of
the 7th Framework Programme (FP7) of the European Commission, addressing (i) the Wireless Sensor Networks and
Cooperating Objects and (ii) the Control of large-scale systems pillars of the Strategic Objective ICT-2009.3.5
"Engineering of Networked Monitoring and Control systems.
CONSERN is introducing a novel paradigm for dedicated, purpose-driven small scale wireless networks characterized
by a service-centric evolutionary approach, referred here as an energy-aware self-growing network and system. In the
context of CONSERN, the Self-Growing network paradigm considers (i) mechanisms for energy efficient interaction of
the wireless network elements and (ii) mechanisms for the reliable and efficient evolvement towards later lifecycle
phases. A Self-Growing network is set up on-demand and is initially dedicated to a single purpose such monitoring
and/or control. Self-growing can be applied on large scale, distributed and cooperative systems, including, for
example, construction sites, and delivering wireless services within a complex home/office environment requiring
network parameter negotiation with a multitude of neighbouring networks, etc. In the course of its lifecycle, it may
coexist and cooperate with other wireless networks of distinct owners and interest groups evolving in the deployment
area using or augmenting existing capacity in order to either serve additional purposes (e.g. in case of an
emergency) or optimising under a specific purpose. CONSERN facilitates the emergence of mechanisms achieving energy
efficient, robust, predictable and self-adaptive operation for future networked systems by considering mechanisms
and methods for autonomic (design- and run-time) energy optimisations as well as cooperative energy aware behaviour
of the distributed system nodes.
Main duties and responsibilities
Project Technical Overview
Analysis, design,
implementation and experimental assessment of novel algorithms energy aware, self growing networks
Participation in the design and implementation of excerpt of the CONSERN prototyping environment. A short demo is provided here:
Self- Management of Cognitive Future Internet
Self-NET (Self-Management of Cognitive Future InterNET Elements - ICT-224344 ) aims to design, develop and validate an
innovative paradigm for cognitive self-managed elements of the
Future Internet. Self-NET shall engineer the Future Internet
based on cognitive behaviour with a high degree of autonomy, by
proposing the operation of self-managed Future Internet elements
around a novel feedback-control cycle. Self-NET shall embed new
management capabilities into network elements in order to take
advantage of the increasing knowledge that characterises the daily
operation of mobile Future Internet users. The consortium shall
develop innovative cross-layer design optimisation approaches that
alleviate the shortcomings and duplication of functionalities in
different protocol layers of the present IP stack. Furthermore,
Self-NET shall provide a peer-to-peer style distribution of
responsibilities among self-governed elements of the Future
Internet, therefore overcoming the barrier of current client-server
and proxy-based models in the operation of mobility management,
broadcast/multicast, and quality of service mechanisms.
Main duties and responsibilities
Analysis, design,
implementation and experimental assessment of novel algorithms for
the self-management of wireless network
nodes
The results of this effort
is depicted in the following publications -1,2
Design and
implementation of the SelfNET prototyping environment. A
short demo is provided
here:
End to End
Efficiency
End-to-End Efficiency (E3 - ICT-216248) is aiming at
integrating cognitive wireless systems in the Beyond 3G (B3G) world,
evolving current and future heterogeneous wireless system
infrastructures into an integrated, scalble and efficiently managed
B3G cognitive system framework from a technical, regulatory,
standardisation and business perspective. The key objective of the
E3 project is to design, develop, prototype and showcase solutions
for optimised usage of existing and future radio access resources.
In particular, more flexible use of frequency spectrum, terminals,
base stations and networks is addressed.
Main duties and responsibilities
Analysis and design of the E3
architecture
The
results of this cross consortium effort are depicted in three papers -1,2,3
Analysis, design, implementation and
experimental assessment of novel architectures for cognitive radio
systems
A standalone prototype designed and implemented by UoA -1
The
integrated prototype available in the end of the project -2
Leader of
sub-workpackage 6.1, "Interfaces,
Integration and Validation of Cognitive
Systems"
End to End Reconfigurability - Phase 2
Building on the successful developments of
the first phase, End-to-End Reconfigurability Phase II (E2R II - IST-2005-027714) will
demonstrate and validate technologies that enable a true seamless
experience based on reconfigurable heterogeneous systems. The
project will pursue research into the most promising directions
towards removing walls (current technical and regulatory
limitations) and building bridges (technical) in order to facilitate
the vision of true end-to-end connectivity. E2R is seen by many
actors of the wireless industry as a core technology to enable the
full potential of Beyond 3G systems. It has the potential to
revolutionize wireless communications just as the PC revolutionized
computing.
Main duties and responsibilities
Analysis and design of the E2R
II architecture
A paper depicting the
actual proof of concept implementation of the
architecture -1
Analysis, design, implementation and
experimental assessment of novel architectures for reconfigurable
networks
A presentation outlining UoA standalone
prototype in the end of the project (presented in the context of E3) - 1
An implementation of reconfigurable
protocols based on the concept of software components - 2
Other excerpts of my work conducted in the context of the project can be found
here,
here,
here and
here
End to End
Reconfigurability
The End-to-End Reconfigurability (E2R - IST-2003-507995)
project aims at bringing full benefits of the valuable diversity
within the Radio Eco-Space, composed of wide range of systems such
as Cellular, Wireless Local Area and Broadcast. The key objective of
the E2R project is to devise, develop and trial architectural design
of reconfigurable devices and supporting system functions to offer
an expanded set of operational choices to the users, applications
and service providers, operators, regulators in the context of
heterogeneous mobile radio systems.
Main duties and responsibilities
Analysis, design, implementation and
experimental assessment of novel architectures for reconfigurable networks
The Reconfiguration Control and Service Provisioning
Manager (RCSPM) platform, designed and implemented by
UoA -1
A UoA lead integrated prototyping
effort, available at the end of the project
The final audit demonstrations, with
UoA participating in all 3 of them, and leading the last
one  -2
Component based, application layer
reconfiguration, part of UoA prototype
Component based, application layer
reconfiguration, part of UoA prototype (longer version - no comments)