Weekly plans and update for week 41

Dear all, we hope all is fine with project 1 close to the deadline. 

Project 2 is now ready if you would like to get started. Just go to the GitHub address of the course, see for example https://compphysics.github.io/MachineLearning/doc/web/course.html and scroll down to project 2.

If you spot typos, errors, inconsistencies etc please let us know asap.  The deadline is set to one month from now, November 9 (Monday). 

Concerning project writing, we remind you that we posted a video early september on how to write projects, see https://www.uio.no/studier/emner/matnat/fys/FYS3150/h20/forelesningsvideoer/WritingScientificReports.mp4?vrtx=view-as-webpage

Feel free to comment it and give us feedback in case something is unclear. We want you write reports which look like a scientific article/report. This is actually one of the learning outcomes of this course. Also, the 

description on how to write the reports and how we evaluate them is at https://github.com/CompPhysics/MachineLearning/blob/master/doc/Projects/EvaluationGrading/EvaluationForm.md

Please, take some time to reread these instructions. Forgetting for example an abstract, penalizes you with  automagic deduction of  5 points out of 100! Make sure you also have proper labels on figures, good figure captions and tables etc.  

This week we will discuss in depth how to write a neural network code (feed forward). We will also discuss how to use Tensorflow and if we get time we start with convolutional NNs. All these lovely topics are in the slides for week 41, see for example https://compphysics.github.io/MachineLearning/doc/pub/week41/html/week41.html

Last week we ended with deriving the back propagation algorithm, together with a discussion on stochastic gradient methods. We start this week by repeating many of these topics. These topics and more are central elements of project 2.  Project 2 starts with you implementing the stochastic gradient method instead of matrix inversion from your your project 1 codes. 

All the best to you all,

Kristian, Michael, Morten, Nicolai, Per-Dimitri, Stian and Øyvind.

ps an example of a report which did very well can be found at https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects/ReportExample