For those of you who don't know Kaggle, it is a portal for Data Science and Machine Learning competitions and resources. Here is the link!
The plan is to start from the simple competitions. They usually have tutorials that explain the most common techniques to approach such problems, but then it is up to you (and the time you want to spend on it) to find better and more efficient solutions.The main point of trying out these problems, is to learn the most used tools and methods in machine learning "by doing the stuff" rather than just learning about the theory.
Once we have tackled some of the easy problems we can start considering the more challenging ones, and depending on how things work, we might get good enough to even participate in those competitions where you can win money!
In general, we will have meetings to discuss together ideas for improving the "state of the art" model, and then work as a team to submit a solution. We will try to work as a real Data Science team in a company, using the meetings to share ideas, and then working on those ideas in smaller groups or individually. We will set up a platform to communicate during the week (e.g. Slack).
The study group may also be useful to get new ideas to discuss in the "Machine Learning" and "Deep Learning" reading groups running in the Complexity Science Department. Or to apply ideas discussed in these groups to Kaggle problems.
Meetings will be held fortnightly on Wednesdays from 4pm to 5pm in Room D1.07, Zeeman Building.
Next meeting will be Wednesday 22 Feb.
Kaggle is kept in very high consideration by companies hiring in the Data Science and the Machine Learning sectors. So, having a good profile there could probably help you find a job after graduation.
More informations will be made available later on, closer to the beginning of term. In the meanwhile, you have the time to familiarise yourself with Kaggle, try out simple competitions and read some blog posts about solutions.
If you are interested, do not hesitate to contact me or Ayman. Also, if you register to Kaggle, please let us know your username, so we can keep in contact on it.
More informations will soon be available on this page.