• International Conference on Electronics, Communications and Computers

Keynotes « CONIELECOMP2017
Image 01

Keynotes

On the adequacy of the Turing Machine Model for Distributed Computations

Dr. Luis Alberto Pineda Cortés

Abstract

The Turing Machine model of computation presupposes that the format of the machine is fully determinate and that information is transformed but not lost in computations. These properties are referred to here as (1) “determinacy” and (2) “information preservation”. In this paper it is argued that distributed formats do not have these properties necessarily and do not conform to Turing’s model for this reason. The claim is illustrated with a diagrammatic computing format in which functions are expressed extensionally and computations are performed by direct operations on such format in a highly distributed fashion. Computing formats that are not fully determinate are entropic and computations have and associated entropy value, which can be defined as the rate of indeterminacy of all computing or memory units involved in a computation. The Turing Machine corresponds to the case in which all units are fully determined and the entropy is zero. The theory is applied to the construction on an associative memory, with the corresponding memory register, recognition and retrieve operations, where contents or “images” are represented as functions and abstractions of such functions. An experiment for the storage and recognition of hand written digits is presented. The experiment shows that there is an entropy value range in which precision and recall are a very high, and that associative memory registers for storing images as attribute-value pairs are very effective, with a very low computational cost. The theory suggests that there is a general computational trade-off between information preservation and expressivity with effective computation. This trade-off suggests in turn that while engineered or artificial computations favor information preservation —but at the cost of the trade-off between expressivity and effective computation or tractability, which is assumed in the Turing Machine model— natural computations favor expressivity with direct but highly distributed computations.

Speaker Biography

Electronics Engineer by Universidad Anáhuac (1976-1980), M. C. in Computing Science by ITESM (1983-1985), Campus Morelos and PhD by the Centre for Cognitive Science at the University of Edinburgh (1986-1990). He was a design engineer at Puromex S. A., (1979-1981) and Data-Center Manager of NCR, Mexico (1981-1983). He started his research career at Instituto de Investigaciones Eléctricas (IIE) where he was a researcher in two occasions (1983-1986) and 1992-1998). He was also a Research Associate at the University of Edinburgh (1989-1992). He joined IIMAS, UNAM in 1998, where he was the Head of Computer Science from 1998 to 2002 and from 2005 to 2010. He is currently a Titular Investigator “B”, PRIDE “D” and SNI-II. He is the author of about 100 research papers, book chapters and international conference proceedings, mainly in Artificial Intelligence, Computational Linguistics, Service Robots, Speech Recognition Natural Language Processing and Computer Graphics. He has developed the research groups SIBS at IIE, and DIME and Golem at IIMAS, UNAM. He has been responsible for several research projects funded by CONACyT, PAPIIT, UNAM and the Institute for Science and Technology of Mexico City. In particular he is the funder and coordinator of the Golem Project that has developed the robots Golem (2002-2008), Golem en Universum (2010), Golem-II+ (2011-2014) and Golem-III. He also started the group for improving the quality of service in the Metro system in Mexico City in 2008, and one of his current PhD students recently created the project to order the access to the trains, with a large practical and media impact.

He has been lecturer at the graduate programs of UNAM, UJAT, Tabasco, CENIDET, Morelos and ITESM, Campus Morelos, in Computing Theory and Logic, Automata Theory and Formal Languages, Programming Languages, Computer Graphics, Computational Linguistics, Artificial Intelligence and Logic Programming. Earlier he lectured on Digital Logical and Computer Architecture at Universidad Anahuac.

He has been invited lecturer in Ohio State where he taught Formal Languages and Automata Theory twice (2003), and also invited professor at Vienna’s Technical University (2011). He has supervised three doctoral dissertations, nine masters and twelve bachelor’s theses, and currently he is supervising five more PhD and two masters theses. Several of his students are now professors and investigators in several institutions and some of them are members of the SNI. Dr. Pineda was also the founder of the doctoral program for Computer Science between the IIE and the University of Salford, UK, in which four researchers at IIE obtained their PhDs. He also participated in the foundation of the M. C. Program in Computer Science at UJAT, Tabasco.

He has given more than 60 seminar talks and conference presentations at prestigious Universities in United States and Europe, and in international events, and has been invited speaker in Mexico more than 80 times. He has been member of the technical committee of several national and international conferences, workshops and seminars. He has also had a large presence in the media (TV, radio, written press and Internet) mainly in relation to the robots Golem, and also in relation to improve the quality of the service of the Metro system in Mexico City.

Dr. Pineda is a regular member of the Academia Mexicana de Ciencias (AMC) and Academia Mexicana de Informática (AMIAC). Was founder and coordinator of Seminario de Computación UNAM (1998-2002), Coordinator of the technical program of the congress 50 Años de la Computación en México in 2008, and also of the Simposium Los Retos de la Era de la Información en nuestro País in 2011.

Dr. Pineda was the coordinator of Red Mexicana de Investigación y Desarrollo en Computación (REMIDEC) from 2010 to 2014. In 2014 coordinated the creation of the Academia Mexicana de Computación (AMEXCOMP) of which he is the founding president. He has organized the 1st and 2nd Reunión Anual de Amexcomp, and he was coordinator and author of the book “Políticas y Estrategias para el Desarrollo de la Computación en México”, which was recently published. He has been a long advocate of the development of computational Technology in México, and has constantly emphasized the role of technology in national sovereignty. He is a marathon runner.

Autonomous Navigation and Exploration with Drones Using Vision

Dr. José Martínez Carranza

Abstract

The drone market is emerging with steady strength. From aerial photography to precision agriculture, drones have become a tool work or even a toy for amusement. However, the arrival of drones to the masses introduces serious implications to be considered, the first one being safety followed by mechanisms to guarantee that non-expert user flies their drones without risk to others. This calls for the development of methods that can run on board the drones, enabling a pilot to receive aid to perform a stable flight, but also for the drone to autonomously decide and execute a safe flight autonomously. Motivated by this, in this talk I will present a set of techniques that we have developed in order to enable drones to carry out autonomous flight by only processing visual information recorded from an onboard camera. No other sensors are used nor GPS or any other information of the environment is known in advanced. Several examples will be presented aiming at illustrating what can be achieved when a drone can see and from this, perform autonomous navigation

.

Speaker Biography

Dr José Martínez-Carranza is an Assistant Professor in the Department of Computer Science and member of the Robotics Laboratory at the Instituto Nacional de Astrofísica, Óptica y Eletrónica (INAOE). He obtained a BSc in Computer Science (Cum Laude) from the Benemérita Universidad Autónoma de Puebla in 2004, and an MSc in Computer Science (Best Student) from INAOE in 2007, both institutions in México. In 2012, he received his PhD from the University of Bristol in the UK, where he also worked as Research Assistant and Associate from 2012 to 2014. His interest includes robotics, computer vision, machine learning applications, and high-performance computing. In the last two years he has received the following distinctions: the highly prestigious Newton Advanced Fellowship (2015-2017), granted by the Royal Society in the UK; In 2015 and 2016, 1st Place in the Mexican Tournament in Robotics, in the service robots @home category; in 2016, Associate Numerary Member of the Mexican Robotics Federation; membership (Level-C 2017-2019) to the National System of Researchers in Mexico; and in 2016, 2nd Place in the International Micro Air Vehicle Competition (IMAV 2016), in the indoors category

.

Towards a thematic network in cyber-physical systems in Mexico: opportunities and challenges

Dr. Gilberto Ochoa Ruiz


Abstract

Cyber-physical systems (CPS) represent a new scientific and technological paradigm, which seeks to respond in a holistic way, to the incessant increase in the complexity of existing infrastructures in all areas of human knowledge. Such systems can be the integration of physical processes, control, and monitoring systems, as well as real-time and intensive computing capabilities. In addition, in general, CPS is a specialization of the “systems of systems” paradigm, which generally involves the interconnection of several components or subsystems through wire or wireless networks, in many cases creating distributed control systems that share many similarities with The Internet of Things.

The economic potential and for society as well as for the development of new technologies is enormous, and many worldwide efforts have been made to develop this area, which requires the convergence of various fields of human knowledge and engineering. This new paradigm requires holistic integration of several sub-disciplines, such as embedded systems in real time, process control and complex process modeling, the study of new communication protocols, among many others.

In Mexico, this field of knowledge remains unexplored and untapped. Today, universities provide their students with highly specialized knowledge, and in a similar way, research centers and laboratories in the country focus their efforts on short-range problems due to technological or competence barriers or lack of knowledge of the activities Carried out by other teams in the country, which prevents the development of multi-disciplinary projects. CONACYT thematic networks seek to improve these problems by supporting national linkage projects with a high scientific and academic impact. For a long time, a group of academics of CONACYT have discussed the creation of a thematic network of Cyber-physical Systems, with a focus on the use of heterogeneous reconfigurable systems, as a way of achieving a paradigm shift in the way in which Engineering is taught in our universities and to encourage the creation of technological innovation projects through joint efforts. This talk will then address the opportunities for the country, as well as the challenges that must be faced to achieve those goals.

.

Speaker Biography

Gilberto Ochoa Ruiz received his PhD degree from the LE2I Laboratory at the University of Burgundy, France, in 2013. Afterwards, he was a post-doctoral research fellow at Lab-STICC laboratory (France) during the period 2014-2015, working in the use of reconfigurable devices and architectures for cyber-physical systems in robotics. Currently, he is a CONACYT Research Fellow at the Mexican Institute of Petroleum, in which he works in the creation of cyber-physical systems, in particular, industrial Process Measurement and Control Systems (IPMCS) based on reconfigurable devices, applied to the field of physical modelling of offshore foundations in large geotechnical centrifuges.

.

Detecting emotions in design process

Dr. Slavko Milekic


Abstract

Emotions play a major role in modern design preferences. Today, using of the shelf consumer technologies, allow designers to take into account potential users’ emotional reaction. These technologies are portable eye- and gaze-tracking as well as portable EEG (brain-wave recordings). This presentation will present an overview of the latest research in the area of “emotional design” done at the University of the Arts in Philadelphia.

.

Speaker Biography

Slavko Milekic holds a medical degree (Belgrade School of Medicine) as well as Ph.D. in Cognitive Science (University of Connecticut). His research interests include digital design, the psychology of human/computer interaction and building of WWW-based tools for knowledge transfer. Dr Milekic’s innovative, child-friendly digital environment (KiddyFace) has been already implemented in a museum setting. He is co-author and principal interface designer for the “Theory of Language”, a CD ROM-based text (MIT Press). Slavko Milekic was granted a US patent for an original way of interacting with a touch-screen display, as well as a patent for using eye- and gaze gestures to interact with digital content.
Dr. Milekic is frequently presenting at international conferences and leading workshops on educational uses of digital technology and innovative methods of human/computer interaction. He has been engaged as a consultant for the development of interactive digital environments by a number of institutions including the Philadelphia Museum of Art, National Constitution Center in Philadelphia, Phoenix Art Museum, Cleveland Museum of Art, Gulf of Maine Aquarium, Daimler & Chrysler Research Laboratory and Interactive Institute in Stockholm, Sweden. He is currently a professor of Cognitive Science & Digital Design at Art + Design Education Department at the University of the Arts in Philadelphia as well as co-founder and Creative Director of Flatworld Interactives, LLC, a company producing innovative software for interacting with visual information. Dr. Milekic’s work has been featured in a PBS series exploring early childhood creativity, “Art to Heart” (www.ket.org/arttoheart/).

.

From Diagnostics to fault tolerance

Dr. Efraín Alcorta García


Abstract

Industrial processes currently have strict safety measures in operation as well as high performance criteria. This motivates the evolution of traditional algorithms dedicated to control towards the development of additional capabilities with which these challenges can be addressed. One of these capabilities can distinguish between a nominal operation (within the tolerance limits of the design) and an operation outside the design nominal values. Also, known as failure operation. In general, a failure can change a value of a parameter outside the established tolerance limits. A failure can affect the system by increasing the energy required to realize the control objective to the weathering of other parts that result in a greater failure, losing the processes of the operation and in some cases, can lead to situations of risk. I noticed a failure can occur in a process for diverse reasons: the weathering of an old part (which can be accelerated by operations being overloaded), failures within the material, etcetera, but are linked to the nature of the processes and there is always a probability of occurrence of failure.
The first step to attenuate and correct the effect of failures is to determine its presence and location within a process. This assignment discusses the conceptuality of a family of methods which has the purpose to detect and locate the presence of failures in a process using the model. The main ideas along with the diverse applications will be demonstrated. In addition, certain types of plants are required to maintain the operation (For example a plane) until it reaches a safe state in which maintenance can be done. In these cases, with the help of the information about the failures it’s possible to build algorithms that allow the continuity of the operations (probably with reduced performance) safely. A set of non-invasive methods, i.e., that do not modify the original connections of the system in which fault tolerance is being performed, are discussed, and shown by examples.

Smart polymers obtained by ionizing radiation: synthesis, characterization, and biomedical applications

Dr. Emilio Bucio Carrillo


Abstract

Different systems of hydrogel copolymers with diverse temperature responsiveness were prepared to elaborate membranes for their biomedical application. The hydrogels were synthesized using thermo-responsive poly(N-isopropylacrylamide) (NIPAAm) and poly(N-vinylcaprolactam) (NVCL), which have a low critical solution temperature (LCST) close to that of the human body temperature and pH-responsive acrylic acid and 4-vinylpyridine. The networks were synthesized using gamma radiation. The low critical solution temperature (LCST) of each system was measured by differential scanning calorimetry (DSC). The effect of using hydrophilic monomers of acrylic acid (AAc), methacrylic acid (MAAc), dimethyl acrylamide (DMAAm), and hydroxyethyl methacrylate (HEMA) for the copolymerization on the critical point was evaluated.

The surface chemistry of grafted samples was analyzed by FTIR-ATR spectroscopy and water contact angle, while their thermal properties were analyzed by TGA and DSC. The stimuli-responsive behavior was studied by swelling and contact angle in water, as well as by DSC. Sensitive films presented a critical pH and LCST.

GE Digital Industrial

Alejandro Herrera, Ramón Reyes and Miguel Camacho






Abstract

The industrial internet and the software platform are expected to be a big market in the coming Years so GE is transforming to a Digital Industrial Company. What does this mean? How GE is changing their Culture? GE is working hard to make Predix and industry standard platform, making its software Open Source. What is Predix?. These questions will be discussed as well as an overview of GEIQ (GE Engineering Center in Mexico) will be provided and how GEIQ is connecting with the Digital Strategy.

Speaker Biography

Alejandro Herrera holds a BS degree in Mechanical Engineering at Universidad de las Americas Puebla. He graduated on May 2007 and one month later he joined GEIQ as a Performance Engineer. During his career, Alejandro has held different roles, from technical to administrative, leading him to his current position as Subsection Manager of the Engine Simulation Technologies team. Among his most outstanding accomplishments in GE are the Engineering Efficiency Recognition for the implementation of an automated system to support engine’s diagnostics, lead the on-wing performance modeling efforts and field support projects such as fuel consumption assessments. Alejandro played basketball with the UDLAP’s team, and he enjoys playing billiard and spending time with his wife and two dogs.

Ramon Reyes holds a BS degree in Mechanical Engineering at Universidad de las Americas Puebla. He graduated on May 2013. From December 2012 to June 2013, he joined GEIQ as an ETIP in the On Point Solutions team. In June 2013 he joined to the Services Performance and analytics team supporting to the creation of predictive models for commercial engines. Among his most outstanding accomplishments in GE are the Engineering excellence award, innovation challenges for Digital visualization and On wing inspection. Ramon likes to play football as a goalkeeper, make trips in bike and have fun with his friends.

Miguel Camacho has a BSc. in Electronics Engineering from the ITT (Instituto Tecnologico de Toluca). He studied his Masters at ITM (Instituto Tecnologico de Monterrey). Miguel worked as Design Engineer in LG Electronics. He joined GE in 2000 as Electrical Systems Engineer. He was Electrical NPI Program Focal in 2005 for M&I (Marine and Industrial). In 2009 he started and led the GEIQ Mechanical Systems Group in GE Aviation Systems. He is Black Belt DFSS, DFR practitioner and PMP certified. He was LM6000 Program Manager for M&I for three Years. He is currently Sub-section Manager for M&I Fluid and Electrical Systems Team. Miguel enjoys being with his family and movies.