Seminarios 2019-2020

Título: Designing IoT systems in the Cloud

Ponente: Dr. Rafael Sánchez

Fechas: 18-19-20 de Mayo de 2020

Horario: de 16:00 a 20:00h

Organizador: Dr. Antonio de Amescua

Lugar: Emisión en directo a través de Google Meet

Créditos: 2 ECTS

Resumen: 

This seminary will cover most of the aspects involved in architecting software for IoT systems and deployments with specific focus on Cloud technologies. The seminary will have a strong use case and a mix of theory and practice; the students will have the opportunity to practice the concepts in a real Cloud sandbox environment.

 

Session 1, May 18th, 16:00h to 20:00h

In Device IoT (systems & software)​

Learning objectives: Understand devices typology and features, OS ecosystem, software to manage device fleets as well as Industry use cases. This session will include labs for the students to do and better grasp the concepts explained.

💻 │ Sesión en directo a través de Google Meet (es necesario logearse con la cuenta de Gmail de la universidad)

 

Session 2, May 19th, 16:00h to 20:00h

IoT Data Architecture

Learning objectives: Have a general understanding of the innards of an IoT Data processing pipeline and the challenges IoT software architectures pose to it. Students will learn about the general architecture, data ingestion and further real time processing. The session will also cover an introduction software testing, delivery and CI/CD processes for IoT devices and backend systems.

💻 │ Sesión en directo a través de Google Meet (es necesario logearse con la cuenta de Gmail de la universidad)

 

Session 3, May 20th, 16:00h to 20:00h

IoT Systems Integration and Data Analysis​

Learning objectives: This session will be focused on IoT Data operations, from Data Analysis and integration with DataWarehouses to Machine Learning modeling from sensor data.

💻 │ Sesión en directo a través de Google Meet (es necesario logearse con la cuenta de Gmail de la universidad)

 

Breve biografía:

Rafael Sánchez holds an Ph.D. in Telematics Engineering (2016) by University Carlos III de Madrid (UC3M) and a 5-year Degree in Telecommunications Engineering (1996) by Polytechnic University of Valencia. Since 1996, he has been involved in multiple enterprise projects in areas like optical networks (SDH/DWDM), IPTV, Ethernet fiber access and networking in companies like Lucent Technologies and Nortel. Currently, he works for Google Cloud as Customer Engineer.

Speaker Personal web page: https://www.linkedin.com/in/rafaelsf80/

Título: Research challenges in privacy measurements

Ponente: Narseo Vallina-Rodriguez 

Fechas y horarios:

• Sesión 1: 31 de enero de 2020, 10:00-13:30 | Sala 4.0.D03
• Sesión 2: 3 de febrero de 2020, 10:00-14:00 | Sala 1.2.C16
• Sesión 3: 4 de febrero de 2020, 10:00-13:20 | Sala 1.2.C16
• Sesión 3: 5 de febrero de 2020, 10:00-14:00 | Sala 1.2.C16

Organizador: Dr. Juan E. Tapiador

Lugar: Leganés, Madrid

Créditos: 

Resumen: 

This seminar will cover fundamentals issues and open problems in the area of Internet measurements focused on privacy research. The contents will cover the following items:

1. Internet measurements 
   1. Active vs. passive measurements
   2. Applications / case studies: traffic analysis, network characterization, security, privacy)
   3. Ethics
2. Privacy
   1. General context of the data economy / digital products - user studies / user perceptions
   2. Regulation: CCPA, COPPA, GDPR
   3. Mobile, web, IoT privacy risks
   4. Cross-device tracking - 3rd parties / SDKs
3. PETS Tools
   1. Tools for privacy analysis and compliance (mobile, web, IoT)
   2. Developing PETS / Platform challenges
   3. Impact of academicresearch and tools
4. Wrap-up
   1. Picking research topics
   2. Open research challenges

Breve biografía:

Narseo Vallina-Rodriguez is an Assistant Research Professor at IMDEA Networks where he leads the Internet Analytics Group. Since 2014, he is also Research Scientist at the Networking and Security team at the International Computer Science Institute (ICSI) in Berkeley. 

Narseo received his degree in Telecommunications Engineering from the University of Oviedo in 2007, which he extended with a 6-month visit at the University of Cambridge to complete his degree dissertation. In 2008, Dr. Vallina-Rodriguez joined Vodafone R&D, returning to the University of Cambridge to complete his Ph.D. program under the supervision of Prof. Jon Crowcroft one year after. In July 2013, Dr. Vallina-Rodriguez joined ICSI in Berkeley (California) as a Post-Doc, becoming a Research Scientist and Principal Investigator one year after.

During his doctoral studies Dr. Vallina-Rodriguez also interned in world-class industry research labs, such as Deutsche Telekom Labs in Berlin (Germany) and the scientific group at Telefonica Research (Spain). The outcome of his research has been awarded with a Qualcomm Innovation Fellowship in 2012, the best short-paper award at ACM CoNEXT'14, he best paper award at ACM HotMiddlebox'15 and a DataTransparencyLabs grant in 2016 for characterizing mobile tracking services with the Lumen Privacy Monitor. Narseo's research has been extensively covered by international media, including Wired, ArsTechnica, NPR, The Verge, ABC Australia, and RTVE (the Spanish public TV) among many others. 

Título: Advances in Artificial Intelligence

Ponentes: Prof. Plamen Angelov, Jamal Berrich y Carlos Rodríguez Pardo

Fechas: 28-29 de Noviembre y 12-13 de Diciembre de 2019

Organizador: Scalab (Jesús García Herrero y Agapito Ledezma)

Créditos: 2 ECTS

Resumen: El seminario está divido en dos partes.

 

PARTE I

Título: Empirical Approach to Machine Learning: How to learn from data (streams) fast, highly accurate and interpretable models

Ponentes: Prof. Plamen Angelov, PhD, DSc, FIEEE

Fechas y horarios:

• Sesión 1: 28 de Noviembre de 2019, de 9 a 12:30h │ Sala Adoración de Miguel (1.2.C16)
• Sesión 2: 29 de Noviembre de 2019, de 9 a 12h │ Aula 4.1.E05

Organizador: José Antonio Iglesias

Lugar: Leganés, Madrid

Créditos: 1 ECTS

Resumen:

We are witnessing an explosion of data (streams) being generated and growing exponentially. Nowadays we carry in our pockets Gigabytes of data in the form of USB flash memory sticks, smartphones, smartwatches etc. Extracting useful information and knowledge from these big data streams is of immense importance for the society, economy and science. Deep Learning quickly become a synonymous of a powerful method to enable items and processes with elements of AI in the sense that it makes possible human like performance in recognising images and speech. However, the currently used methods for deep learning which are based on neural networks (recurrent, belief, etc.) is opaque (not transparent), requires huge amount of training data and computing power (hours of training using GPUs), is offline and its online versions based on reinforcement learning has no proven convergence, does not guarantee same result for the same input (lacks repeatability).

The lectures will introduce a new concept of empirical approach to machine learning which has proven convergence and local optimality for a class of such models. In this course, the newly introduced Empirical Approach to Machine Learning will be presented in a systematic way. Its applications to problems like anomaly detection, clustering, classification, prediction and control will be illustrated. The major advantages of this new paradigm is the liberation from the restrictive and often unrealistic assumptions and requirements concerning the nature of the data (random, deterministic, fuzzy), the need to formulate and assume a priori the type of distribution models, membership functions, the independence of the individual data observations, their large (theoretically infinite) number, etc.

Referencias:

[1] P. Angelov, X. Gu, Empirical Approach to Machine Learning, Studies in Computational
Intelligence, vol.800, ISBN 978-3-030-02383-6, Springer, Cham, Switzerland, 2018.

[2] P. P. Angelov, X. Gu, Deep rule-based classifier with human-level performance and
characteristics, Information Sciences, vol. 463-464, pp.196-213, Oct. 2018.

[3] P. Angelov, X. Gu, J. Principe, Autonomous learning multi-model systems from data
streams, IEEE Transactions on Fuzzy Systems, 26(4): 2213-2224, Aug. 2018.

[4] P. Angelov, X. Gu, J. Principe, A generalized methodology for data analysis, IEEE Transactions on Cybernetics, 48(10): 2981-2993, Oct 2018.

[5] P. Angelov, Autonomous Learning Systems: From Data Streams to Knowledge in Real time, John Willey
and Sons, Dec.2012, ISBN: 978-1-1199-5152-0.

 

Session 1. 28 November 2019, 9am-12:30

Introduction to the Empirical Machine Learning ​

Learning objectives: To introduce the students to the concepts of the Empirical approach to Machine Learning, to familiarise with the basics such as recursive density estimation, typicality, anomaly detection, clustering, classification and to compare with the traditional approach. To be ready to get from experimental data to empirically derived models. 

 

Session 2. 29 November, 9am-12:00

Autonomous Learning Multi-MOdel (ALMMo) Systems and Deep Rule-based (DRB) classifiers and their applications​

Learning objectives: To introduce ALMMo and DRB and their design and applications to solve classification and prediction problems. To provide a range of examples including image processing. The learning can be facilitated if the students also access the freely available lecture notes to my book as well as software (Matlab and Python code) at https://angeloventelsensys.wixsite.com/plamenangelov/software-downloads.

 

Breve biografía:

Prof. Angelov (MEng 1989, PhD 1993, DSc 2015; IEEE Fellow, 2015; Fellow IET) is Vice President of the International Neural Networks Society (INNS). The supervisor of his PhD, Dr. Dimitar P. Filev is a Member of the USA National Academy of Engineering. Prof. Angelov holds a Personal Chair in Intelligent Systems at Lancaster University, UK.

He has authored or co-authored 300 peer-reviewed publications in leading journals, peer-reviewed conference proceedings, 3 granted patents, 3 research monographs cited 8800+ times with an h-index of 48 and i10-index of 156. His single most cited paper has 940+ citations. He has an active research portfolio in the area of computational intelligence and machine learning and internationally recognised results into online and evolving learning and algorithms for knowledge extraction in the form of human-intelligible AI.

Prof. Angelov leads numerous projects (including several multi-million ones) funded by UK research councils, EU, industry, UK MoD. His research was recognised by ‘The Engineer Innovation and Technology 2008 Special Award’ and ‘For outstanding Services’ (2013) by IEEE and INNS. He is also the founding co-Editor-in-Chief of Springer’s journal on Evolving Systems and Associate Editor (AE) of several leading international scientific journals, including IEEE Transactions on Cybernetics, IEEE Transactions on Fuzzy Systems, Soft Computing, etc. He is also Area Editor of the Fuzzy Sets and Systems journal and served as AE for Applied Soft Computing journal (2013-2018). He gave 25 keynote talks at high profile conferences and co-organised and co-chaired 30 IEEE conferences including the high profile IJCNN2013 held in Dallas, TX, IJCNN2015 held in Killarney, Ireland, the INNS Conference on Big Data (the inaugural one held in San Francisco in 2015 as well as a series of annual IEEE Conferences on Evolving and Adaptive Intelligent Systems.

Prof. Angelov was International Programme Committee Chair for a number of top IEEE Conferences such as FUZZ-IEEE (2014, 2018), IJCNN (2016, 2019), Intelligent Systems (2014, 2016), CYBCONF (on Cybernetics) 2013, etc. Dr Angelov is the founding Chair of the Technical Committee on Evolving Intelligent Systems, System, Man and Cybernetics Society of the IEEE (a Governor of which he also was during the term 2015-2017) and was previously chairing the Standards Committee of the Computational Intelligent Society of the IEEE (2010-2012). He was also a member of International Program Committee of over 100 international conferences (primarily IEEE). 

Personal web page: http://www.lancaster.ac.uk/staff/angelov

 

PARTE II

Título: Agent oriented programming and simulation/visualization in real-world applications

Ponentes: Jamal BERRICH y Jorge López Moreno

Fechas y horario:

• Sesión 1: 12 de Diciembre de 2019, de 9 a 14h
• Sesión 2: 13 de Diciembre de 2019, de 9 a 12h
• Sesión 3: 13 de Diciembre de 2019, de 12 a 14h

Organizadores: Jesús García Herrero y Miguel A. Patricio

Lugar:  Sala Adoración de Miguel (1.2.C16). Campus de Leganés

Créditos: 1 ECTS

Resumen:

The seminars will present some paradigms and tools coming from artifical intelligence and virtual reality applied yo business examples.

In the first part of seminars, the Agent-Oriented Programming (AOP) methodology will be presented and applied to real examples, including a platform to illustrate some examples. AOP is a new programming paradigm that enables the provision of autonomous, active objects with artificial intelligence for the business and functional processing of a given application. The application process of this paradigm is guided by modeling, design and development methodologies and is managed by MAS platforms for agent deployment. The agent which represents the essential component of a multi-agent system is subjected to types of architecture including the BDI architecture. In order to realize an agent-oriented application, the developer finds himself in the embarrassment of choosing the type of the design methodology and the type of SMA platform of the deployment. This task becomes very embarrassing in the absence of a clear approach in light of the conventional software engineering method. In this perspective, we propose a new approach of software design taking for example the joco platform by offering a means managing the interoperability between the platforms SMA.

In the second part of seminar, the experience of a spin-off company will be presented, describing the technologies based on simulation and visualization to improve the user experience and commercialization strategies. Seddi is a company whose main field of action is in the creation and development of cutting-edge technologies in the field of textile simulation. Born as spin-off of Universidad Rey Juan Carlos, it is focused in simulation tools and 3D visualization, appliying these techniques under research towards the textile sector. Among the technologies developed by Saddi, we can highlight mechanical simulation, acquisition of mechanical parameters, flat and knitted fabric models, mechanical simulation at thread scale, optical appearance capture and hyperrealistic simulation of thread scale. These technologies allow to offer solutions in the design and testing phases, within the textile sector. It can also be a novel element both in the field of online marketing and commercialization, and in the transition from the fashion industry to a digital model that generates a more reliable shopping experience for the user.

 

Session 1, 12 December, 9:00 – 14:00

Agent oriented programming from theory to practice​

Learning objectives: defining MAS system and how to use it

 

Session 2, 13 December, 9:00 – 12:00

AOP practice with BDI architecture

Learning objectives: how to use BDI architecture to develop a business application: Joco platform

 

Session 3, 13 December, 12:00 – 14:00

The case study of SEDDI​

Learning objectives: how to apply simulation and visualization techniques in the sector of fashion manufacturinga and commercialization. A business success case

 

Breve biografía:

Jamal Berrich

Jamal BERRICH is assistant professor in an engineering school called ENSAO (National School of Applied Sciences Oujda, Morocco). He is also member of research team SIQL in the LSE2I Lab (Laboratoire des Systèmes Electroniques, Informatique et Images). He works as software engineer and developer and currently he is working on Artificial Intelligence field. Among others, he is currently working on an Artificial Intelligence project of civilian drone called 'SKYNET project' in partnership with an industrial start-up ATLAN Space.

• https://www.linkedin.com/in/jamal-berrich-08310132/
• https://www.youtube.com/channel/UCpN6Yf9HJuPHZQlgntbGpYQ/videos

Jorge López Moreno

He is one of the creators of the mechanical and optical fashion simulation technology that has allowed the creation of DESILICO. He obtained his PhD in Computer Engineering under the supervision of Diego Gutiérrez and Erik Reinhard (U. Bristol, UK), receiving the Extraordinary Thesis Award from the University of Zaragoza. During his studies, funded by Adobe Systems, he spent more than 8 months at the headquarters of Adobe Advanced Technology Labs (San Jose, California) and obtained three patents in image processing techniques and 3D editing. After two years of postdoctoral stay at INRIA (France), he receives twice the Juan de La Cierva national scholarship as a researcher at the Rey Juan Carlos University.

It has tens of publications in JCR conferences and magazines of maximum international relevance, such as: ACM Transactions on Graphics, IEEE TVCG or Computer Graphics Forum. He has experience in creating start-ups (Omepet, WorldPathol) and technology transfer and patents with leading design and 3D graphics companies: Adobe Systems, Activision, Autodesk, Next Limit, Solid Angle, etc.