Plataforma GAT: Información traslacional biomédica para tratamientos alternativos de cáncer.

Es química pura de profesión con maestría en ciencias, especializada en química computacional y ciencia de datos. Actualmente trabaja en el diseño y desarrollo de una plataforma de Investigación Traslacional para la investigación de cáncer en el Programa de Generación de Alternativas Terapéuticas en cáncer GAT en la Pontificia Universidad Javeriana de Bogotá. En esta misma universidad es docente del programa de Ciencias de Datos en el Departamento de Ciencia de la Información de la Facultad de Comunicación y Lenguaje. Es cofundadora de las comunidades Modelamiento Molecular Colombia y la Red de Mujeres en Bioinformática y Ciencia de Datos de Latinoamérica, y líder de la comunidad de PyLadies Colombia.

Pronóstico de las precipitaciones mensuales vinculadas a sus teleconexiones con los índices de gran escala – Un enfoque amplio para su implementación con Redes Neuronales Artificiales

Wilfredo Alfonso-Morales, Ingeniero Electrónico (2007), Magíster en Ingeniería (2010), Doctor en Ingeniería (2016), y Especialista en Estadística Aplicada (2020); todos sus títulos en la Universidad del Valle. Se desempeña actualmente como profesor asistente en la Escuela de Ingeniería Eléctrica y Electrónica de la Universidad del Valle. Recientemente categorizado como investigador Senior por el Ministerio de Ciencia y Tecnología. Es miembro del grupo de Percepción y Sistemas Inteligentes – PSI. Autor de varios artículos y un libro. Su trabajo se ha realizado en temáticas relacionadas con la Inteligencia Computacional aplicada a diferentes áreas del conocimiento en Ingeniería (Ambiental, Sanitaria, Topográfica, Civil, Industrial, Eléctrica y Electrónica) y  la Salud (Medicina, Odontología y Rehabilitación). Dentro de sus temas de investigación se encuentran: el aprendizaje automático y profundo, los algoritmos evolutivos, la analítica de datos, el control inteligente, entre otros.

AIoT – When IoT meets the AI

Marcelo Rovai

Microcontrollers (or MCUs) are very cheap electronic components, usually with just a few kilobytes of RAM, designed to use tiny amounts of energy. They can be found in almost any consumer, medical, automotive, and industrial device. It is estimated that over 40 billion microcontrollers will be sold this year, and probably there are hundreds of billions of them in service nowadays. These devices don’t get much attention because they’re often only used to replace the functionality of older electro-mechanical systems in cars, washing machines, or remote controls.  More recently, with the Internet of Things (IoT) era, a significant part of those MCUs is generating “quintillions” of data, that in its majority is not used due to the high cost and complexity (bandwidth and latency) of data transmission.  In recent decades, we have seen a lot of development in Machine Learning models trained with huge amounts of data in very powerful and power-hungry mainframes. What is happening today is that it is now possible to take noisy signals like images, audio, or accelerometers and extract meaning from them by using neural networks. And what is more important is that we can run these networks on microcontrollers and sensors themselves using very little power, interpreting much more of those sensor data that we are currently ignoring. This is TinyML, a new technology that enables machine intelligence right next to the physical world. We believe that TinyML can have many interesting applications for the benefit of society at large.

Marcelo Rovai was born in São Paulo and held a Master’s degree in Data Science from the Universidad del Desarrollo (UDD) in Chile and an MBA from IBMEC (INSPER) in Brazil. He graduated in 1982 as an Engineer from UNIFEI with a specialization from Escola Politécnica de Engenharia of Universidade de São Paulo (USP); both institutions are located in Brazil. 

Mr. Rovai has experience as a teacher, engineer, and executive in several technology companies such as CDT/ETEP, AVIBRAS Aeroespacial, SID Informática, ATT-GIS, and NCR, DELL, COMPAQ (HP), and more recently at IGT as a VP. He now works at IGT as a Senior Advisor for Latin America.

Besides his work at IGT, Marcelo Rovai publishes articles about electronics on websites such as MJRoBot.org, Hackster.io, Instructables.com, and Medium.com. Furthermore, he is a volunteer Professor at the UNIFEI Engineering Institute in Brazil and a lecturer at several Congresses and Universities on the topics of IoT and TinyML. He is an active member and a Co-Chair of the TinyML4D group, an initiative to bring TinyML education to developing countries.

Data Centric AI un Medicine

Jose David Posada

«El campo de la medicina no ha sido ajeno a los recientes avances en el campo de la inteligencia artificial. El campo ha visto como se han implementado métodos novedosos para mejorar el rendimiento en aplicaciones de diagnóstico, estratificación de riesgo y extracción de información a partir de datos no estructurados entre muchas otras. No obstante, todas estas aplicaciones dependen en su mayoría de contar no sólo con acceso a información médica que de por si puede llegar a ser difícil debido a asuntos regulatorios, sino que dicha información se encuentre en un formato adecuado para responder las preguntas de investigación. Incluso si esto pasa, en casos donde se hace necesario utilizar información de varios instituciones simulatáneamente es bastante común que cada una tenga un forma diferente de representar la informaicón y por ende de evaluar su contenido. En esta charla nos centraremos en la inteligencia artificial centrada en los datos o Data Centric AI , disciplina que nos lleva a concentrar los esfuerzos en mejorar la información que tenemos para construir mejores modelos utilizando técnicas de Intliegencia Artificial.»

Jose David Posada es PhD en Biomedical Informatics de la Universidad de Pittsburgh. Previo a su posición actual como Profesor Asistente en el departamento de Ingeniería de Sistemas de la Universidad del Norte estuvo en la Universidad de Stanford como Científico Senior de Datos Clínicos ayudando a construir la segunda generación del repositorio de datos estandarizados para la investigación STARR-OMOP (https://arxiv.org/abs/2003.10534 ). Es el líder del grupo de trabajo para latinoámerica de OHDSI (https://ohdsi.org). Se especializa en la estandarización de registros médicos electrónicos y procesamiento de lenguaje natural con texto clínico. Web page: www.posadajose.com twitter: @posadajd

Thinking Outside the CMOS Box
Dr. Victor Grimblatt – R&D Group Director and General Manager, Synopsys

Even if the progression of Moore’s law continues relentlessly, in 2040 it would still take the world fastest supercomputer a billion years to factor RSA-2048, and the simulation of nitrogenase enzyme co-factor, a fairly simple organic molecule, would not be completed before the end of our universe.

The world is dramatically changing; we are facing problems that were unimaginable last century. Global warming triggering droughts and natural disasters, social claims related to inequality, pandemics, unemployment because of Industry 4.0, etc. We are living a complicated time that is putting us face to new and different challenges that technology should help to solve. And we can ask ourselves, do we have the right hardware and computational power for these challenges?

The computing and memory requirements of artificial intelligence (AI), biochemistry, medicine, pharmacology, and physics applications greatly exceed the capabilities of current electronics and are unlikely to be met by isolated improvements in devices, or integrated circuit architectures alone.

Circuits based on super-conducting electronics (SCE) have made many advances in the last few years and are the basis of Quantum Computers (QC). Exascale computing demands for exascale bandwidth: silicon photonics is capable of dramatically faster data communications, at a much lower power.

Discrete and monolithic 3D-IC will allow the full potential of heterogeneous integration of different functions and may boost their independent evolution, both w.r.t. the process technology and the design tools.

This talk offers an overview of the rising innovations that may change our industry and our future.

Processing-in-memory (PIM)-based Manycore Architecture for Training Graph Neural Networks
Partha Pratim Pande, FIEEE (https://eecs.wsu.edu/~pande/)

I am a Professor at the school of EECS, Washington State University (WSU), Pullman, Washington. Effective August 2013, I am also the holder of the Boeing Centennial Chair in Computer Engineering. I joined WSU in 2005, as an Assistant Professor. I received my PhD in Electrical and Computer Engineering from the University of British Columbia 2005. I obtained my M.S. in Computer Science from the National University of Singapore in 2002 and B.Tech in Electronics and Communication Engineering from University of Calcutta, India in 1997. My CV can be found here.

An Introduction to Quantum Computing and Quantum Hardware

Victor Rodriguez-Toro received his BS degree in Electronics Engineering at Universidad del Valle in Cali, Colombia. After an exchange program as an Erasmus Mundus scholar at the Technical University of Munich (TUM) in Munich, Germany, he received his MS degree in Electronics Engineering from Universidad del Valle. Awarded with a Fulbright Fellowship, he obtained his MS and PhD degrees in Electrical and Computer Engineering (ECE) from the Georgia Institute of Technology in Atlanta, GA. During his graduate studies, Dr. Rodriguez-Toro was recognized by the School of ECE as an
Outstanding Graduate Teaching Assistant and given the ECE Faculty Award, and by the College of Engineering with the Intel FOCUS scholar award. Currently, Dr. Rodriguez-Toro works at IBM Research as a Hardware Developer in Quantum Computing in Yorktown Heights, NY. His previous research experiences include the study of molecules by using electronic structure and quantum dynamics numerical packages, and the modeling of nanostructures/materials by using semiconductor, electromagnetic, and RF software
tools. Dr. Rodriguez-Toro also has research/industry experience in the design, fabrication, and/or characterization of electronic/photonic/quantum devices and circuits for applications in human-computer interaction, ubiquitous computing, biomedical sensing, clean energy harvesting, virtual/augmented reality, storage, and communications. His current research interests also include the areas of Quantum
Information Science and Technology based on superconducting materials. Dr. Rodriguez-Toro has been an instructor and mentor in STEAM initiatives in Colombia and the US, and he has served on scientific committees of conferences in solar energy in Latin America.