XX machine learning spain meetup y tenemos una programación de lujo para celebrarlo!! Esta es la agenda: 18:30 – Acceso al Auditorio de Campus Madrid. 18:40 – «Mapping Poverty with Machine Learning and Satellite Imagery». In developing countries that need the most resources for eliminating poverty, data on poverty is particularly scarce due to the high cost of conducting surveys. This data gap hampers efforts to study and design policies for reducing poverty. Stanford’s Sustainability and Artificial Intelligence Lab has been working towards closing this poverty data gap by estimating consumption expenditure and asset wealth from publicly available high-resolution satellite imagery. Our initial work focused on five African countries: Nigeria, Tanzania, Uganda, Malawi, and Rwanda. First, we trained a convolutional neural network to learn image features that are useful for predicting nighttime light intensities, a rough proxy for economic wealth. We then optimized the neural network by training on limited survey data. After this two-step process known as transfer learning, the convolutional neural network is able to identify image features which explain up to 75% of the variation in local-level economic outcomes. Finally, we implemented a machine learning pipeline for automating the production of global-scale poverty maps using any dataset or model. This pipeline represents a first step towards creating up-to-date poverty maps to guide nonprofit organizations and policymakers. Ponente: • Christopher Yeh Christopher Yeh is a junior undergraduate at Stanford University and studies computer science with a concentration in artificial intelligence. A native of California, he is currently studying abroad in Madrid’s International Institute for the spring quarter, learning about Spanish culture and improving his Spanish fluency. He has always been very interested in environmental sustainability. Since joining Stanford’s Sustainability and Artificial Intelligence Lab in 2016, he has helped work on the poverty mapping project, developing a scalable machine learning pipeline for estimating poverty levels in developing countries. In the past, he has interned at Intuit and Apple, working on iOS app development. Outside of academia, he enjoys volleyball, photography, outdoor activities, and playing the cello. 19:30 – «Hidden Markov Models: nuevas aplicaciones para viejos trucos«. Quien lleve suficiente tiempo atento a los avances en inteligencia artificial recordará que, antes del resurgimiento de las redes neuronales en los últimos años, otras técnicas fueron responsables de muchos avances, y entre ellas HMM ocupó un lugar destacado, sobre todo en lo que relativo al reconocimiento de voz. En esta charla descubriremos cómo, lejos de ser cosa del pasado, HMM sigue siendo una herramienta de machine learning muy relevante, con importantes avances teóricos y nuevas aplicaciones. Ponente: • José Luis Hidalgo Después de muchos años trabajando en multinacionales de consultoría (Accenture) y telecomunicaciones (Ericsson y Huawei), Jose Luis decidió dar el salto al mundo de las start-ups de la mano de Nextail Labs, una empresa española que está revolucionando la logística en el sectore de la moda a base de usar técnicas de machine learning y optimization operativa. 20:20 – ¡Cervezas y networking! Tomaremos unas cervezas y podremos seguir charlando sobre los temas de interés que salgan en esta jornada. Nos vemos el martes 13 de junio en Campus Madrid! ]]>