Plenary Talk : Meta-Learning : Beyond Fitting a Model to Data
Machine learning algorithms build a mathematical model of training data describing a problem at hand. There are many approaches which model to choose and how to set its parameters. Because the range of machine learning techniques is very wide and heterogenous, it can be difficult to choose a suitable ona for a particular problem. The meta-learning approach utilizes previous experience with learning algorithms to improve the results of machine learning applications. By applying machine learning principles to machine learning process itself it is possible to automate tasks that have so far been dependent on human expertise and experience. In the talk we will present several examples of meta-learning techniques to optimize neural network models, and to propose complete data mining workflows.
Roman Neruda is with the Institute of Computer Science of the Czech Academy of Sciences, department of machine learning, where he is working in the areas of neurocomputing, evolutionary algorithms and meta-learning. He graduated from the Faculty of Mathematics and Physics, Charles University, and obtained his CSc degree in the ICS CAS. In 1995-1996 he was with the Los Alamos National Laboratory, he has been working on joint project with colleagues from Carnegie-Mellon University, Koblenz Universitaet, University of California Chico, University of St. Etienne, and Universidad Distrital Bogota. He is the co-author of more than a hundred international publications. He teaches evolutionary algorithms and multi-agent systems at Faculty of Mathematics and Physics, Charles University
Plenary Talk: The Automated Data Science and Automated Machine Learning Revolution
Abstract. In recent years, the demand for Data Science (DS) and Machine Learning (ML) experts has outpaced the supply. To address this gap, there has been an increasing number of of automated user-friendly DS andML tools that can be used by non-experts.In this talk we will give a quick review of DS and ML. We will examine IBM Watson Analytic, no longer available, but considered the first automated DS and ML tool which aim was to allow final users to explore insights and learn models form data by just asking questions in natural language. Then we overview three of the most popular available AutoDS and AutoML tools: DataRobot, TPOT.and H20 AutoML and conclude reflecting on the possible future this ongoing revolution.
Watson (computer) https://en.
This is Jeopardy! on 12/05/2013 YouTube https://www.youtube.
IBM’s Watson Supercomputer Destroys Humans in Jeopardy YouTube https://www.youtube.
IBM Watson is a technology platform that uses natural language processing and machine learning to reveal insights from large amounts of unstructured data http://www.chooseportal.
TPOT: A Python Tool for Automating Data Science https://www.kdnuggets.
Genetic Algorithm in Artificial Intelligence – The Math of Intelligence https://www.
Google launches an end-to-end AI platform https://techcrunch.
German Hernandez is an Associate Professor at the Department of Systems and Computing Engineering at the National University of Colombia, Bogota, Colombia. He leads there the Research Group in Algorithms and Combinatorics “Algos” and the Research Group in Algorithmic Trading, Computational Finance “Algotrade”. German is also a partner of Algocodex, an algorithmic trading company. He has a PhD. in Math with Conc. in CS and an MSc. in CS form U of Memphis; an MSc, in Math., an MSc. in Stat. and BSc in Computing and Systems Eng. form the National University of Colombia. He also has been an adjunct professor at the CS Dept at U of Memphis.
Plenary Talk: Autonomous Vehicles a Reality with 5G
Vehicle connectivity requires real-time data transmission, low latency and high effectiveness. This conference presents the benefits of 5G networks for autonomous vehicles, their interaction with other areas such as artificial intelligence and data analytics, integrating aspects of security and flexibility. The future of connected vehicles in Latin America and the opportunities for monetization for the different actors of the ecosystem will be discussed.
Ericsson Cluster North, Colombia.
Cayo Betancourt has more than 18 years of professional experience in the telecommunications sector. He obtained his degree at the Universidad del Cauca and an MBA from the University of Liverpool. He joined Ericsson in 2001, and is currently responsible for coordinating, directing and alienating an extensive digital transformation program. Cayo actively contributes to LinkledIn by sharing his knowledge and experiences in the development and implementation of new technologies, including also issues of leadership and organizational culture. Cayo has significant experience in operations in America, Asia and Oceania..