Con motivo de la tercera edición de Big Data Coruña, hablamos con Paco Nathan, uno de los comunicadores más importantes en el campo del Big Data.

En este evento nos contó la experiencia de transformación sufrida por O’Reilly Media en la adopción de Machine Learning en sistemas de producción y la metodología en torno a Notebooks.

1. What does Big Data mean in you career?
   Big Data began in the late 1990s, in Silicon Valley, although it become more widespread in the middle 2000s after cloud was introduced. For me, that was mid-career. On the one hand, early in my career I had studied about statistics, optimization, machine learning, AI — though more practically I had also studied systems engineering and software architecture. The latter was much more important to employers initially. However, after Big Data practices became introduced, then both parts of my background become valuable together. In other words, Big Data catalyzed a real change in my career, and opened many more opportunities.
2. What would you highlight about Big Data Coruña?
   There is excellent expertise represented here, both in the latest frontiers of machine learning as well as in practical applications in business. I think that combination is essential for progress. It is a hallmark of these past three years of Big Data Coruña.
3. How does this event contribute to the academic sector? And to the business one?

   This continues from the previous question and answer. I find it interesting in Silicon Valley that there is much less separation now between the academic sector and the business one. Fei-Fei Li, one of the leading AI researchers at Google — arguably the AI leader in industry — she also leads AI research at Stanford University.  Similarly, my former boss Ion Stoicha was CEO of Databricks and now he is the chairman of EECS at UC Berkeley, leading RISElab, which in turn has much participation from industry leaders in machine intelligence. 

   So I think it is more important than ever, now, for maintaining a balance between academic progress and business applications. This event contributes to both, again with a very good balance.
    

4. Which are the main developments of this edition?

   There have been industry applications of Big Data, Machine Learning, Artificial Intelligence, which are becoming ubiquitous: how we drive in cars, how we communicate with smartphones, etc. This edition of Big Data Coruña comes at a good time to reflect on those advances, to understand their significance. What impact does that have for university students looking to join machine intelligence projects in industry?
    

5. What would you highlight about your papers on Machine Learning’s adoption in production systems and the methodology concerning Notebooks?

   Both have a common theme in what is called “Human-in-the-loop” as a design pattern for organizations. The idea is that AI does not replace people; instead, people who have experience in a business can contribute more by becoming augmented through AI.  In machine learning, we have generally called this Active Learning.

   For example here is a talk about human-in-the-loop by Eric Colson, who leads machine learning at the fashion company StitchFix.  AI is used to suggest apparel for customers, with experts assisting where the machines need “advice”.

   I think it is crucial that we do not think so much about AI “replacing” people, rather that people use technology to accomplish more.
    

6. What takes you to collaborate with this event?

   I am always grateful for opportunities to visit A Coruña and my friends here at the university. They are doing fantastic work, and we are their fans at O’Reilly Media.  Also, I really appreciate the opportunities here to discuss with other experts in this field — while enjoying the wonderful setting of A Coruña.