Monica Vitali Curriculum Vitae Personal Information Name Monica Vitali Date and Palermo, April 16, 1984 Place of Birth Address Phone E-Mail Web Page via G. Ponzio 34/5, 20133 Milan, Italy (+39) 02 23993424 monica.vitali@polimi.it home.deib.polimi.it/vitali/ Academic Positions 01–2014/now Research Assistant (Assegnista di Ricerca), Politecnico di Milano, Milano, ITALY. Research in the field of Information Systems with a focus on adaptive systems and the assessment and management of energy efficiency in applications, in collaboration with the Information System group under the supervision of Prof. Barbara Pernici. 01–2011/12– Ph.D. Candidate, Politecnico di Milano, Milano, ITALY. 2013 Research in the field of adaptive systems for improving efficiency of applications in data centres using machine learning techniques. 10–2010/10– Research Assistant (Assegnista di Ricerca), Politecnico di Milano, Milano, 2012 ITALY. Research in the field of data centres energy efficiency, assessed through the definition and implementation of metrics at the infrastructure and application level. 07/2009– Research Assistant (Collaboratore alla Ricerca), Università degli Studi 06/2010 di Palermo, Palermo, ITALY. Research in the field of robotics and artificial intelligence with a focus on imitation learning models. Work Experience and Training 03–09/2008 Software Developer, S.I.R.I.O., Palermo, ITALY. Implementation of a wireless sensor network for vineyard and historical monuments monitoring related to two regional projects: Virtus Vini and Geomon T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 1/9 03–06/2008 Intern, National Research Council (CNR), Palermo, ITALY. Research in self-organizing systems; development of a Jade/Prolog interface; development of a Jade based agent society in which individuals share Prolog knowledge 03–07/2006 Intern, National Research Council (CNR), Palermo, ITALY. Research in multi-agent systems and genetic programming: development of a Java/Jade application using genetic programming techniques in agent systems Education 2011–2013 Ph.D in Information Technology (XXVI cycle), Politecnico di Milano, Milan. Thesis: “Measuring and Improving Energy Efficiency of a Data Center in a Self-Adaptive Context”, advisor Prof. Barbara Pernici. 09/2012– Visiting Student, Massachusetts Institute of Technology, Cambridge, 02/2013 USA, Supported by the Rocca Foundation scholarship. Study about machine learning techniques for Bayesian Networks automatic learning and probabilistic operator selection. 2006–2009 Master of Science in Computer Science for Intelligent Systems (Laurea Specialistica in Ingegneria Informatica per i Sistemi Intelligenti), Università degli Studi di Palermo, Palermo, First-Class Honours (110/110 e Lode). Thesis: “Study of an Agent Model Based on the Genetic Metaphor”, advisor Prof. Salvatore Gaglio. 2003–2006 Bachelor of Science in Computer Science (Laurea in Ingegneria Informatica), Università degli Studi di Palermo, Palermo, First-Class Honours (110/110 e Lode). Thesis: “Genetic Programming in Multi-Agent Systems: an Application”, advisor Prof. Salvatore Gaglio. Research Statement I am a research assistant at Politecnico di Milano. I am working in the Information Systems group, under the supervision of Prof. Barbara Pernici. The focus of my current research is on Energy Efficiency and on Green Information Technology at the application level. I am exploring methods and models to reduce the impact of data centers and clouds on the environment and to reduce CO2 emissions while saving energy and money. In parallel to this research I am exploring the field of Artificial Intelligence. I used different techniques in the last years and I consider them as the basis of every system that claims to be adaptive. In my research I am connecting these two topics (Green IT and Artificial Intelligence) to create an adaptive system for managing Energy Efficiency. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 2/9 Current research focus Green IT My research is mainly focused on managing energy efficiency and sustainability and Green of applications and services. In literature, most of the approaches focus on the Cloud exploitation of the hardware efficiency and on server consolidation. These approaches try to propose solutions which are independent from the application nature but they do not take into account how the application behaviour can dynamically change over time. In my research I propose a comprehensive approach which takes into consideration the behaviour of the application in a dynamic environment. This goal is reached through several steps. The first step consists in the definition of the business goals. In this way it is possible to define the constraints in terms of quality and sustainability that the application has to respect. The second step consists in the monitoring of the application. This is realized through the selection of relevant metrics able to catch the behaviour of the application and in the definition of new metrics. Once metrics are defined, the application state has to be assessed by monitoring the relevant metrics and verifying the satisfaction of constraints. The last step consists in the adaptation enactment. If some constraints are not satisfied, adaptation actions are selected from a pool of available strategies in order to solve or mitigate the violation. The definition of a proper set of actions and the selection mechanism are object of the research. In the specific, selecting the proper action given a specific context can be trivial with an extended number of variables involved and the effect of actions can be positive or negative for several of the involved variable. An automatic mechanism for action selection is needed and has to be based of a continuous refinement process for learning and updating the knowledge about the effect of actions over the monitored variables. Metrics and Through the definition and computation of specific metrics it is possible to understand Monitoring the behaviour of an application and to detect issues related to quality and efficiency. The selection of proper metrics is essential for successfully managing the application. In my research I am focusing on the existing metrics and also proposing new metrics for describing the application in terms of energy efficiency and sustainability. I am investigating hidden relations between metrics. The knowledge of these relations enables a deeper understanding of the application and of the environment in which it has been deployed. This acquired knowledge can be used for predicting the effect of a modification and for conducting “what if” analysis. Finally, I am investigating techniques for detecting the metrics which better describe the application efficiency and quality for specific applications and for class of applications (data intensive, CPU intensive, memory intensive). Metrics are computed starting from data retrieved by a monitoring system. The amount of information to be collected can be too big for big data centers or cloud environment. The collection of such an amount of data can generate inefficiencies due to delays introduced by the monitoring system. For these reason I am investigating techniques for reducing the amount of data to be monitored. I am also studying distributed techniques for collecting and analysing data in order to improve the collection mechanism and to reduce the time needed to conduct analysis on the monitoring data. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 3/9 Adaptive Adaptation is the key for achieving efficiency and quality in dynamic environment Systems such as the management of applications in data centers and clouds. In my research I am proposing adaptive algorithms and techniques for reducing energy consumption of applications and maintaining their quality of service. The algorithms are based on the informed selection of a repair strategy from a set of available strategies. The selection is based on the learned and continuously updated knowledge about the ability of the action of fixing or mitigating the observed inefficiency. Adaptive Ambient Assisted Living Recently I am investigating the issue of ambient assisted leaving to provide an adaptive environment in order to assist elderly and non self-sufficient subjects. This kind of systems requires the employment of a set of sensors and a monitoring system to detect interesting events occurring in the observed environment. The aim of the research consists in generalizing the models proposed for the Green IT adaptation to this new field and to provide mechanisms for efficiently collecting and analysing relevant data. The elaboration of data coupled with the knowledge of the habits of the observed individuals and some reference values for a set of metrics can be the trigger for a set of adaptation actions allowing the management of emergencies or simply used for helping the individual in his daily activities. Machine learning techniques are employed to associate adaptation actions to the detected events, in order to better fit the specific individual life-style and to adapt to the modifications of his habits. Scientific background My early research, developed within the Computer Science department at the University of Palermo, was focussed on artificial intelligence techniques for machine learning and adaptation. I have developed a demonstration-driven learning algorithm for teaching an agent how to solve simple skills observing a set of demonstrations where the problem was solved by a human. The goal was to discover recurrent patterns in the demonstrations for identifying sub-goals of the problem to solve, decomposing the whole problem into a set of simpler ones without any explicit knowledge. Generalization techniques have also been applied to learn general rules from experience. I have also investigated cooperation between intelligent agents by using evolutionary algorithms. The aim of the research was to start with a population of agents unable to solve a specific problem and to evolve the population towards a better set of agents with the necessary skills. The work required the formalization of a genome inspired description of the agent abilities and on the development of techniques for creating new individuals by crossing two genomes. I have also developed pattern recognition techniques applied to digitized images in order to discover the presence in the images of scratches or noise. These experiences gave me the background for working with self-adaptive systems using machine learning techniques in my current research. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 4/9 Research Activities Participation in International Research Projects 10/2012– Eco2 Clouds, Experimental Awareness of CO2 in Federated Cloud Sourcing, 7th 09/2014 Framework European Project, http://www.eco2clouds.polimi.it/. Development of methods, guidelines and technology for enriching cloud computing with means to take into proper consideration ecological concerns, such as energy consumption and CO2 footprint of applications. 10/2010– GAMES, Green Active Management of Energy in IT Service centres, 7th Framework 06/2012 European Project, http://www.green-datacenters.eu/. Monitoring and management of energy efficiency in a data center using adaptive techniques to improve applications efficiency while preserving the quality of service. 05/2019– HUMANOBS, Humanoids that Learn Socio-Communicative Skills by Observation, 09/2010 7th Framework European Project, http://www.humanobs.org/. Development of new cognitive architectural principles to allow intelligent agents to learn skills by observing and imitating the behaviour of a human demonstrator. Participation in National Research Projects 04/2014–now Attiv@bili, Assistenza digitale e innovazione Sociale in Ambienti Controllati, Bando Smart Cities Regione Lombardia. Analysis and evolution of adaptive processes for social and health care addressed to elderly and not self-sufficient people. Analysis and management of monitoring data in an Ambient Intelligence environment. 03/2008– VirtusVini, Vineyard Monitoring through Wireless Sensor Networks, Regione Sicilia, 09/2008 http://www.dicgim.unipa.it/networks/wsn/projects/virtusvini.html. Implementation and management of a wireless sensor network for monitoring environmental features in a vineyard with the aim of enhancing the quality of the production and reducing risks of failures. 03/2008– GeoMon, Structural Health Monitoring of Historical Heritage Buildings, Regione Si09/2008 cilia, http://www.dicgim.unipa.it/networks/wsn/projects/geomon.html. Design and implementation of an innovative technological framework for monitoring critical structures and geological sites employing a wireless sensor network. Program Committee Service – Program Committee member of the 2014 IEEE International Conference on Systems, Man, and Cybernetics (SMC2014) Research collaborations – Prof. Una-May O’Reilly and Dr. Kalyan Veeramachaneni of the Massachusetts Institute of Technology, Cambridge - USA: joint research in the field of machine learning techniques applied to adaptive management of energy efficiency in data centres [J1][C4]. – Prof. Salvatore Gaglio, Dr Massimo Cossentino, and Dr Riccardo Rizzo of the National Research Council (ICAR-CNR), Palermo - IT: joint research in the field of evolutionary programming for coordination of multi-agent systems [C9][C11]. – Prof. Antonio Chella and Prof. Haris Dindo of the Università degli Studi di Palermo, Palermo - IT: joint research in the field of learning by imitation techniques [C8]. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 5/9 List of Attended Conferences and Conventions 2013 Participation to the international conference on Systems, Man, and Cybernetics, and presentation of the paper “Modeling Service Execution on Data Centers for Energy Efficiency and Quality of Service Monitoring” [C4], October 13-16 2013, Manchester, UK 2012 Attendance of the DataCentres Europe 2012 industry-oriented conference and presentation of a Demo related to the GAMES European project, May 23-24 2012, Nice, France Attendance of the 2012 VMware Academic Research Symposium and presentation of the poster “Resource Allocation in Virtual Machines for Energy Efficient Data Centers”, September 26-27 2012, Cambridge, Massachusetts, USA 2011 Participation to the SIGMETRICS international conference and presentation of the paper “Usage centric green performance indicators” [J4] at the GreenMetrics Workshop, June 7-11 2011, San Jose, California Participation to the international conference on Intelligent Computer Communication and Processing (ICCP) and presentation of the paper “Business process co-design for energy-aware adaptation” [C7], August 24-26 2011, Cluj-Napoca, Romania Teaching Activities AY Contract Professor, 30 hours, Sistemi Informativi per il Settore dell’Informazione, 2014/2015 Politecnico di Milano (COMO). Teaching Assistant, 20 hours, Sistemi Informativi per il Settore dell’Informazione (Prof. Barbara Pernici), Politecnico di Milano. Teaching Assistant, 20 hours, Sistemi Informativi (Prof. Cinzia Cappiello), Politecnico di Milano. Teaching Assistant, 40 hours, Fondamenti di Informatica (Prof. Maristella Matera), Politecnico di Milano. AY Teaching Assistant, 20 hours, Sistemi Informativi per il Settore dell’Informazione 2013/2014 (Prof. Barbara Pernici), Politecnico di Milano. Teaching Assistant, 20 hours, Sistemi Informativi (Prof. Cinzia Cappiello), Politecnico di Milano. Teaching Assistant, 24 hours, Fondamenti di Informatica (Prof. Maristella Matera), Politecnico di Milano. AY Teaching Assistant, 20 hours, Sistemi Informativi per il Settore dell’Informazione 2012/2013 (Prof. Barbara Pernici), Politecnico di Milano. AY Teaching Assistant, 20 hours, Sistemi Informativi per il Settore dell’Informazione 2011/2012 (Prof. Barbara Pernici), Politecnico di Milano. Teaching Assistant, 24 hours, Fondamenti di Informatica (Prof. Maristella Matera), Politecnico di Milano. Tutoring, 20 hours, Matlab (Scuola di Ingegneria dell’Informazione), Politecnico di Milano. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 6/9 Awards and Recognitions 2012 Progetto Roberto Rocca Fellowship, a scholarship for promoting joint research between MIT and Politecnico di Milano as a visiting Ph.D. Candidate (Autumn Term 2012) 2012 Scholarship for participating to the 2nd 804 COST Training School on Energy Efficiency in Large Scale Distributed Systems held in Palma de Mallorca, 24-27 April 2012 2012–2013 PhD Scholarship awarded by the Ministry of Education, University, and Research (MIUR), Italy Publications International Journals [J1] Monica Vitali, Barbara Pernici, and Una-May O’Reilly. “Learning a goal-oriented model for energy efficient adaptive applications in data centers”. In: Information Sciences (accepted manuscript available online since 7 February 2015), pp. 1–19. doi: 10 . 1016 / j . ins . 2015 . 01 . 023. url: http : / / www . sciencedirect . com / science / article / pii / S0020025515000614. [J2] Monica Vitali and Barbara Pernici. “A Survey on Energy Efficiency in Information Systems”. In: International Journal of Cooperative Information Systems (IJCIS) 23.03 (2014). doi: 10.1142/S0218843014500014. [J3] A. Kipp et al. “Applying green metrics to optimise the energy consumption footprint of IT service centres”. In: International Journal of Space-Based and Situated Computing (IJSSC) 2.3 (2012), pp. 158–174. doi: 10.1504/IJSSC.2012.048897. [J4] Doron Chen et al. “Usage centric green performance indicators”. In: ACM SIGMETRICS Performance Evaluation Review, Presented at ACM Greenmetrics International Workshop, San Jose, California, USA, May 2011 39.3 (2011), pp. 92–96. doi: 10.1145/2160803. 2160868. Refereed Conferences and Workshops [C1] Monica Vitali. “Managing Energy Efficiency and Quality of Service in Cloud Applications Using a Distributed Monitoring System”. In: 23rd Italian Symposium on Advanced Database Systems (SEBD), Gaeta, Italy, June 14-17, 2015. Springer LNCS, 2015 (accepted for publication), pp. 1–12. [C2] Cinzia Cappiello et al. “Sustainable choices for cloud applications: a focus on CO2 emissions”. In: ICT for Sustainability 2014 (ICT4S-14), Stockholm, Sweden, August 25, 2014. Atlantis Press, 2014, pp. 352–358. doi: 10.2991/ict4s-14.2014.43. [C3] Cinzia Cappiello et al. “Eco-reports in Clouds”. In: Joint Workshop Proceedings of the 2nd International Conference on ICT for Sustainability 2014, Stockholm, Sweden, August 24-27, 2014. CEUR-WS.org, 2014, pp. 10–13. [C4] Monica Vitali, Una-May O’Reilly, and Kalyan Veeramachaneni. “Modeling Service Execution on Data Centers for Energy Efficiency and Quality of Service Monitoring”. In: IEEE International Conference on Systems, Man, and Cybernetics, Manchester, SMC 2013, United Kingdom, October 13-16, 2013. IEEE Computer Society. IEEE, 2013, pp. 103–108. doi: 10.1109/SMC.2013.25. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 7/9 [C5] Cinzia Cappiello, Pierluigi Plebani, and Monica Vitali. “Energy-Aware Process Design Optimization”. In: EuroEcoDC workshop at the 2013 International Conference on Cloud and Green Computing, Karlsruhe, Germany, September 30 - October 2, 2013. IEEE. IEEE Computer Society, 2013, pp. 451–458. doi: 10.1109/CGC.2013.77. [C6] Barbara Pernici et al. “Setting energy efficiency goals in data centers: the GAMES approach”. In: Workshop on Energy Efficient Data Centers (E2DC). Vol. 7396. 2012, pp. 1–12. doi: 10.1007/978-3-642-33645-4_1. [C7] Cinzia Cappiello et al. “Business process co-design for energy-aware adaptation”. In: Intelligent Computer Communication and Processing (ICCP), 2011 IEEE International Conference on. IEEE. 2011, pp. 463–470. [C8] Haris Dindo et al. “Learning problem solving skills from demonstration: an architectural approach”. In: Artificial General Intelligence - 4th International Conference, AGI 2011, Mountain View, CA, USA, August 3-6, 2011. Proceedings. Vol. 6830. Springer LNCS, 2011, pp. 194–203. doi: 10.1007/978-3-642-22887-2_20. [C9] Monica Vitali et al. “A Genome based Vision of Multi-agent Systems”. In: ICAART 2011 Proceedings of the 3rd International Conference on Agents and Artificial Intelligence, Rome, Italy, January 28-30, 2011. SciTePress, 2011, pp. 406–409. isbn: 978-989-8425-41-6. [C10] E Ardizzone et al. “Multi-directional detection of scratches in digitized images”. In: Proceedings of the 17th European Signal Processing Conference (EUSIPCO’09). European Signal Processing Conference. European Association for Signal Processing, 2009, pp. 248–252. [C11] A. Ciuro et al. “Towards a New Inheritance Definition in Multi-Agent Systems”. In: Proceeding of Workshop on Objects and Agents (WOA 2008). Seneca Edizioni, Nov. 2008, pp. 54–60. isbn: 978-88-6122-122-2. Submitted Manuscripts [S1] Monica Vitali. “Analysing Energy Efficiency of HPC Applications in a Data Center through an Ontology-Driven Model”. In: Computer Science - Research and Development (CSRD), Springer (2015 (under revision)), pp. 1–8. [S2] Cinzia Cappiello et al. “CO2 -aware Adaptation Strategies for Cloud Applications”. In: IEEE Transactions on Cloud Computing (IEEE TCC) (2015 (under major revision)), pp. 1–14. [S3] Usman Wajid et al. “On Achieving Energy Efficiency and Reducing CO2 Footprint in Cloud Computing”. In: IEEE Transactions on Cloud Computing (IEEE TCC) (2015 (under major revision)), pp. 1–14. Reports and Thesis [R1] Monica Vitali. “Measuring and improving energy efficiency of a data center in a self adaptive context”. PhD thesis. Politecnico di Milano, Italy, 1–222, 2014, pp. 1–222. url: https://www.politesi.polimi.it/handle/10589/89864. [R2] Gareth Francis, Kostas Kavoussanakis, Monica Vitali, et al. WP2 Requirements and Overall Architecture: D2.4 Updated ECO2 Clouds Architecture. Tech. rep. ECO2 Clouds 7th Framework Programme Project, 2014. [R3] Gareth Francis, Kostas Kavoussanakis, Monica Vitali, et al. WP4 Deployment Optimisation and Adaptation Algorithms: D4.3 Optimization and Adaptation of Application Deployment in Federated Clouds (II). Tech. rep. ECO2 Clouds 7th Framework Programme Project, 2014. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 8/9 [R4] Barbara Pernici, Cinzia Cappiello, Monica Vitali, et al. Evolution Methodology, Deliverable D3.4 WP3. Tech. rep. GAMES 7th Framework Programme Project, 2010. Opensource Released Tools dcSimulation, A tool for simulating monitoring data in a data center by managing load rate and defining features of the involved servers and applications, https: //github.com/monicavit164/dcSimulation. StructureLearning, A tool for learning relations existing between a set of continuous variables represented using a Bayesian Network, https://github.com/ monicavit164/StructureLearning. StructureLearning, A tool for training and testing an adaptive action selection algorithm for suggesting the best repair action given the state of a set of variables and constraints defined upon them, https://github.com/monicavit164/ ActionSelection. T (+39) 02 23993424 • B monica.vitali@polimi.it Í home.deib.polimi.it/vitali/ 9/9