Our speakers

Keynote

Cloud AutoML

In this session we will learn together how to build on Google's machine learning experience to create our own tailored custom machine learning models. We will see how you can take your own data to train, evaluate, improve and deploy models to meet your business needs in areas such as natural language or vision.

Lykeion, one ML platform to rule them all

The daily basis of a data scientist consists of analyzing data, testing hypotheses and contrasting them, and finally giving the solution to the different parties involved in the process. However, this last mile of getting in production the models created is not an easy task. In this talk, we are to show you the approach that we follow in Cabify to make it easier for data scientists to put into value their work to the different parties through Lykeion, our own ML production-ready platform.

TinyML: The next wave in AI

Over the past few years, we have witnessed how AI has improved at an unprecedented speed. Improvements in processor speed and the arrival of big data have revolutionized the field. Models such as the recently released GPT-3 and Turing-NLG have achieved impressive results that raised awareness of the capabilities of AI. These models follow the current trend of building bigger models to achieve better results. However, developing a bigger model does not always imply a better model. Factors such as computational resources, memory and power to train and run models play an essential role in future of AI. Tiny machine learning is a discipline emerging in the field that focuses on building Machine learning solutions baring in mind these crucial factors. TinyML has the potential to transform many industries. In the session, we will introduce TinyML; we will dive deep into how it works, how to implement it, as well as some interesting use cases.

AI: Think globally, act locally

Artificial Intelligence is the beginning of a paradigm shift that has been predicted for decades. But how can we, as individuals, be part of the revolution? In this session we will discuss the steps we need to take to benefit from AI today, bringing the global thought to a local acts. Let’s discover together the paths to successfully implement AI in our business (and which steps to avoid) through real examples of applied AI.

Democratizing AI to empower individuals, organizations and society

This is an unprecedented moment in human history. Technologies are emerging and affecting our lives in the Fourth Industrial Revolution in new and unanticipated ways. This is a new era that builds and extends the impact of digitization. Artificial intelligence is not new to us. It’s only in recent years that AI has made such impressive progress, primarily driven by an exponential increase in computing power, vast amounts of data, and notable breakthroughs in algorithms. We will experience intelligence in every facet of our lives: in the products and services we use, how we communicate and relate to each other, how organizations function and collaborate, and how society and countries evolve. With such enormous potential, it is no wonder that all enterprises are ramping up their investments in AI. However, there are challenges. Everyone acknowledges that AI has broad transformative potential, but enterprises struggle to translate this potential into tangible advantages. Some of the key obstacles to organizations’ progress with AI include the requirements to establish a strategy and goals, justify and secure investments for projects. Without a comprehensive strategy, enterprises often utilize AI only in limited instances. These single-use cases only scratch the surface of the potential of AI. The real power of AI is in its ability to holistically transform the enterprise and redefine business in ways beyond our present frames of reference

Los 7 puntos en común de los experimentos altamente efectivos

El experimento es el primer paso hacia una modelo de machine learning exitoso. En esta charla, explicaremos siete buenas prácticas para la planificación y ejecución de un experimento de machine learning que nos permita llegar a la consecución de un producto de alta calidad.

From Data Warehouse to Data Lakehouse

In this session, we will embark on a Data Strategy journey that will take us from the old and venerable Data Warehouse of the 80s to the modern data platforms based on data lakes, and the data lakehouse concept. Using the history of data platforms as an excuse, and our very own Sidra Data Platform implementation as a test subject, we will learn the "Dos and Don'ts" and more importantly, we will understand the state of the art of Modern Data Platform designs.

Ejemplos de aplicación de IA en la industria

En esta charla nos introduciremos en el mundo de los algoritmos genéticos, por qué son atractivos, que fundamentos los motivan y qué oportunidades ofrecen. Presentaremos un par de casos de usos reales en un entorno industrial (job-shop-scheduling) y en uno de consumo (gestión de la demanda).

How to embed ethical, explainable, diverse and inclusive AI in AI ecosystem

As computers behave more like humans, how will they impact real people? Skilling up for an AI-powered future requires more than science, technology, engineering, and math. As computers behave more like humans, the social sciences and humanities will become more important. The development and deployment of AI must be guided by the creation of an ethical framework. There are six core principles that should guide the work around AI. Four core principles of fairness, reliability & safety, privacy & security, and inclusiveness underpinned by two foundational principles of transparency and accountability. These six principles are necessary but not sufficient and they do not answer every question. One big question that we have been pondering is whether we might see a Hippocratic oath for coders, similar to the Hippocratic oath that doctors take that says that first, they will ‘do no harm’ – The Future Computed. Ultimately, this discussion around ethics and codes of practice looks at how we as a society can stay true to timeless values in developing and delivering this new technology. Angela will discuss how to promote human-centered, transparent, and ethical AI by design.

Seamless MLOps with Seldon and Mlflow

Deploying and managing machine learning models at scale introduces new complexities. Fortunately, there are tools that simplify this process. In this session, we walk you through an end-to-end hands-on example showing how you can go from research to production without much complexity by leveraging the Seldon Core and MLflow frameworks. We will train a set of ML models, and we will showcase a simple way to deploy them to a Kubernetes cluster through sophisticated deployment methods, including canary deployments, shadow deployments and we’ll touch upon richer ML graphs such as explainer deployments.

Fighting Phishing with Deep Learning and Image Recognition

Cyber attacks are becoming a serious issue for most companies, losing millions due to constant business disruptions. What once was just a problem for a few, has now become a challenge for everyone, from large and small companies to personal email accounts. Phishing is one of the main entry points for most of those cyber threats we face today, being used to deliver all kinds of malware or to trick employees into sharing sensitive information. In this talk, we will show how we can use Deep Learning and image recognition to detect malicious URLs that try to look like legitimate sites so that employees will perceive them as trustworthy and type in their credentials.

Inteligencia Artificial. Reinventando la tradición

La Ingeniería es uno de los sectores donde el conocimiento de las personas es un valor añadido que marca la diferencia la mayoría de las veces. Este conocimiento, basado en la tradición pero con un fuerte contenido tecnológico, se complementa a la perfección con las nuevas tecnologías que han aparecido en los últimos años. Todo ello lleva a que estemos viviendo una auténtica revolución en la ingeniería, que lleva a hacer las cosas de una manera diferente en muchas ocasiones pero conservando la esencia de siempre. Más que peder el sabor tradicional, sin lugar a dudas, lo estamos reinventando.

Level up your game using Azure and Reinforcement Learning!

When you hear about machine learning, you hear about different approaches, mainly two: supervised learning and unsupervised learning. However, there’s a third approach that is associated often with gaming that is called reinforcement learning. During this session, we are going to discover what reinforcement learning is, how it differs from the other approaches to machine learning, and how to implement a reinforcement learning model on Azure. How? Can we make a Minecraft agent walk safely around a maze? What do you think, are you ready to learn?

Transformando la atención médica mediante Deep Learning

Esta charla nos ofrece un recorrido a través de los últimos avances en el análisis de datos médicos utilizando modelos basados en Deep Learning, haciendo especial énfasis en lo referente al diagnóstico por imagen. Se revisarán también los principales problemas a los que se enfrenta un desarrollador de herramientas de Inteligencia Artificial dentro de este contexto. 

Potenciar la aplicación de estándares a través de IA, el camino de Aenor.

"Los estándares disponen una gran cantidad de información que abarca a los diferentes sectores productivos. Hasta ahora, su consumo se basaba casi exclusivamente en el uso de ficheros PDF, requiriendo de una curva de aprendizaje elevada para su uso y su aplicación. Durante la charla contaremos la experiencia de Aenor utilizando IA, no solo para buscar información de forma conceptual, sino también para realizar un descubrimiento de entidades como fórmulas, definiciones, reglas, ensayos, … para su aplicación en los procesos productivos de las empresas, además de ser la base para la ideación de nuevos modelos de negocio dentro de Aenor. Discutiremos también sobre la estrategia de implantación de IA en Aenor para potenciar tareas que hasta ahora eran exclusivamente manuales o estaban fuera del alcance de los sistemas clásicos."