Last week I participated the Aalto Digi Breakfast on Data Science. The Digi Breakfast meetings cover timely topics in digitalization in Aalto University, and the aim is to better connect the world-class research and companies in the Helsinki area.
The program was quite tight, with a lot of short and even shorter talks from both Aalto and companies. There were lots of familiar faces from my time in the Aalto university and HIIT, and it was interesting to now listen to the talks from the company side.
The first keynote was given by Hekki Mannila, a famous computer scientist and current president of the Academy of Finland. He gave the following ten suggestions for data scientists (emphasis mine):
- Think big - look for really big opportunities
- Think small - start with easy wins
- Do not follow the herd - explore new problem domains
- Talk to people in other camps
- Work on a few things
- Do not assume that only size matters - small data can be more important than big data
- Beware and take advantage of technological development
- Simplify - small models often work best
- Assume infinite computational power
- Have fun!
I really liked these, especially the ones about not following the herd! Through the open knowledge community I have had the opportunity to talk people from a wide range of background, which as been really eye-opening and also produced a lot of data science application ideas.
Next Johanna Bragge and Jukka Nurminen presented resent developments in the data science education in Aalto University. These included a new minor in analytics and data science, and ideas for a data science project course and/or hackathon (such as this), where companies could provide some interesting data for the students to play with. Here I suggested also using open data that could be used to study important societal challenges.
In the research spotlights from Aalto, Sami Kaski (my Phd supervisor) summarised the research conducted in the Centre of Excellence in Computational Inference Research (COIN). The focus of COIN is to create both faster inference methods for statistical models and models that scale to really large data sets.
Keijo Heljanko introduced Aalto Data Hub, a networking group for data science researchers and companies. Concrete ideas were to share interesting data sets and also experiences in using big data tools. The open source bigdata cluster we have been developing at Avaus should be relevant here.
Ville Peltola talked about creative discovery and especially the ever-interesting IBM Watson, “cognitive computing system”. Watson has been now applied to cooking, and an API is also provided for developers. Nice!
Also a couple of companies gave short talks. CSC - IT Center for Science is building data science services, and their Pouta cloud could be a good alternative for Amazon with data stored in Finland. Other companies present including Comptel, ZenRobotics, Reaktor and Elisa. They boasted with their data science expertise and/or expressed interests in research collaboration.
All in all, a very good session in just one and a half hours. Will look forward to the future sessions and maybe even presenting something related to open data or my data, who knows.