MeetUp Data Science Northeast Netherlands: June 22, 2017
Mydatafactory wants to bring together starting and experienced developers and researchers in the area of data science in an informal setting.
A great opportunity to share ideas, experiences and to meet the Mydatafactory team and (core) technology!
18:00 - 18:30 Pizzas, drinks & networking
18:30 - 19:00 Bulls-eye: infinite ways of observing cows (Ingo Wassink)
19:00 - 19:30 Ethnic profiling in the 21st century (Victor de Graaff)
19:30 - 21:00 More drinks & networking
Abstracts and bio's can be found below.
Bulls-eye: infinite ways of observing cows - Ingo Wassink, Nedap
Nowadays, it is normal to have smart watches and cell phones to keep track of human health status. These systems know sooner and more accurate what your current health status is and can help you to improve your health. At Nedap Livestock Management, we are developing similar systems, not for humans, but for cows. Our systems continuously observe millions of cows by collecting sensor data, to farmers about their cattle. What data are needed to keep track about cow’s health status and how can the data be collected and managed?
Ingo Wassink did his masters and PhD at the Computer Science department at the University of Twente. Now he is active in the research & development at Nedap Livestock Management. He is responsible amongst others for the Nedap Velos platform. He will talk about how this platform is used to collect & analyze sensor data to keep track of animal health status.
Ethnic profiling in the 21st century - Victor de Graaff, independent consultant @ ING
A spark of negativity hit the Dutch news recently when it became common knowledge that the Dutch police forces use a technique called ethnic profiling; decisions to pull over cars were made based on easy to identify indicators such as race, age, car brand, and the price of their car. These indicators are created by people, and scored based on prior success rates by humans. While many of such decisions are shifting towards computer systems, it is the responsibility of the data scientist to ensure that such stereotypes are not shifted along with them. In machine learning, we can define this as a two-fold problem: (1) a bias in feature creation, and (2) a bias in training data. Whereas the first one in the above example is obvious in the race feature, the second problem may be slightly more hidden in the car brand feature. In this talk, we will discuss how to cope with both of these problems, and how we will sometimes have to settle for an imperfect system to not contribute to an imperfect society.
Victor de Graaff is a freelance data scientist, currently working at the fraud & cybersecurity department of ING Bank. He received his Bachelor degree in electrical engineering at the TU Delft in 2008, and a Master degree in computer science at the University of Twente in 2009. After a short period in the industry, he successfully defended his PhD thesis on Geosocial Recommender Systems at the University of Twente in 2015.