Beskjeder
Dear all, first a great thank you to all of you for heroic efforts with the various projects throughout the semester.. We just wanted to update you about the lab this week and next week.
Tomorrow, Tuesday November 28, we have two lab sessions, 1015am-12pm and 215pm-4pm.
We will have at our disposal room F?465 at the CCSE (our regular room F?434 is occupied tomorrow).
On Wednesday we are back in F?434 and the lab is open from 10am to 4pm.
Next week we will have a lab session on Wednesday only, December 6 from 10am-4pm in F?434 (our regular lab room).
We hope this can be of help for you all with the finalization of project 3. We plan also to upload our feedback to project 2 as soon as possible (we are aiming at the end of this week but some of the feedbacks may be delayed since some of the TAs are busy with their own exams).
Wishing you all the very best and please don't hesitate to contact us,
Adam, Daniel, F...
Dear all, thx again for heroic efforts with project 2. We really appreciate this, it is a true pleasure to see what you folks have achieved.
Else, there is an error in the weekly schedule on the official UiO site. We have a regular lab session tomorrow and Wednesday this week, as well as next week (week 48). And possibly week 49 as well. I will send out more info later today about the schedule this week. Else, our last lecture is Thursday November 23.
We have thus lab tomorrow Tuesday November 21 from 815am-4pm and Wednesday as well. The topic is obviously a discussion of project 3.
We have our last lecture this coming Thursday with a discussion of random forests and boosting methods, before we wrap our course.
Next week, week 48, we plan to have lab sessions Tuesdays and Wednesday 10am-4pm.
We are aiming at doing this during week 49 as well, but probably with fewer hours since this is an exam week....
Dear all, we are very impressed by both project 1 and the work you are doing on project 2. It is fantastic to be at the lab and discuss with you. Here we add three references which could be of interest when writing the report and also for further studiers.
The first reference is a preprint where the authors have tested different activation functions. It can be useful when you are writing your report. Furthermore, from arXiv you can fetch the bibtex code as well as the source file for the article by clicking on the other formats link (upper right corner).
The link to this article is at https://arxiv.org/abs/2109.14545
There are also two books on the mathematics of deep learning which are very interesting, one longer at
https://arxiv.org/abs/2310.20360 with codes at https://github.com/introdeeplearning/book (thx to August Femtehjell) and a shorter book at
https://arxiv.org/abs/2105.04026
These two books contain a discussion of the mathematics...
Dear all, this week and the next week are our two last weeks this semester and we will discuss decision trees, random forests and ensemble methods like Boosting and Bagging. Our plans are as follows
Active learning sessions on Tuesday and Wednesday
- Work and Discussion of project 2
- Discussion of project 3 as well, project 3 is now available at https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects/2023/Project3
Material for the lecture on Thursday November 16, 2023
- Thursday: Basics of decision trees, classification and regression algorithms and ensemble models like random forests and Bagging
- Readings and Videos:
- Lecture notes at https://github.com/CompPhysics/MachineLearning/blob/master/doc/LectureNotes/week46.ipynb
- ...
Dear all, we hope the weekend has elapsed in the best possible way for you all. Last week we were unfortunate since our lecture hall was reserved for high school teachers (annual event across the university). We ended thus up with project work only and had to defer the lecture on convolutional neural networks to a recording only. I am however not happy with the recording I made this weekend and will thus make a new one on convolutional neural networks (CNNs). Hopefully it will be done by the end of Monday November 6.
However, till then you can watch the recordings on CNNs from last year at https://youtu.be/MdYT6uwOkT0 (second lecture) and https://youtu.be/3bDkrB-E7cU.
This week we end our discussions on deep learning methods by presenting another method, namely recurrent neural networks. We...
Dear all and welcome back to a new week!
This week we start with convolutional neural networks. This topic continues next week with recurrent neural networks as well.
That will end our discussion of deep learning methods. The final methods we will discuss and study are decision trees, random forests, bagging and boosting methods. The latter methods will be discussed during weeks 46 and 47 (our last weeks).
This week we plan to
Material for the active learning sessions on Tuesday and Wednesday
- Exercise on writing your own neural network code, application to the OR and XOR gates, see notes from last week. Same exercises this week
- The exercise this week is a continuation from last week
- Discussion of project 2
- Video of lab session from l...
Dear all, here are our plans for week 43:
Plans for week 43
Material for the active learning sessions on Tuesday and Wednesday.
Exercise on writing your own neural network code, application to the OR and XOR gates
The exercises this week will be continued next week as well
Discussion of project 2
Material for the lecture on Thursday October 26, 2023.
Building our own Feed-forward Neural Network and discussion of project 2, continuation from last week
Solving differential equations with Neural Networks and intro to Tensorflow with examples.
Readings and Videos:
These lecture notes, see attachment below
...
Dear all, two short messages first during this hectic weekend with project 1 and the exercises for week 41.
1) You can hand in the exercises for week 41 with deadline Sunday 15 on Sunday the 22nd, and you can hand them in together with the exercises for week 42. You will get a score for both then. If you were busy with project 1 till late today, feel free to hand this in later.
2) If you need to adjust project 1, feel free to do so till tomorrow at the end of the working day, that is Monday october 16 at 6pm in the afternoon. We will not subtract points. We will start grading after that. GitHub/Gitlab have a time stamp. Thus, feel free to amend/correct if it got too hectic today!
Best wishes to you all with the project.
Else, the plans for this coming week are
Plan for week 42
Mater...
Dear all, we hope you have had a great weekend. The plans for this week are (although the deadline for project 1 has been extended to October 15)
Plans for week 41
Active learning sessions on Tuesday and Wednesday
- Exercise on writing your own stochastic gradient and gradient descent codes. This exercise continues next week with studies of automatic differentiation and will be used in project 2
- One lecture at the beginning of each session on the material from weeks 39 and 40 and how to write your own gradient descent code
- Discussion of project 2
- Your task before the sessions: revisit the material from weeks 39 and 40 and in particular the material from week 40 on stochastic gradient descent
Material for the lecture on Thursday October 12, 2023
- Neural Network...
Plans for week 40
- Work on project 1 and discussions on how to structure your report
- No weekly exercises for week 40, project work only
- Video on how to write scientific reports recorded during one of the lab sessions
- A general guideline can be found at https://github.com/CompPhysics/MachineLearning/blob/master/doc/Projects/EvaluationGrading/EvaluationForm.md.
- Lecture October 5
- Stochastic Gradient descent with examples and automatic differentiation
- Neural Networks, setting up the basic steps, from the simple perceptron model to the multi-layer perceptron model.
- Readings and Videos:
- ...
Dear all and welcome to a new week with FYS-STK3155/4155. We hope you've had a pleasant weekend.
The plans for this week are tentatively as follows:
Material for the active learning sessions on Tuesday and Wednesday
- Discussions on how to structure your report for the first project
- Exercise for week 39 on how to write the abstract and the introduction of the report and how to include references. See link below.
- Link to the exercise
- Work on project 1, in particular resampling methods like cross-validation and bootstrap. For more discussions of project 1, chapter 5 of Goodfellow et al is a good read, in particular sections 5.1-5.5 and 5.7-5.11.
These sec...
Dear all, we hope you are all doing well and had a great weekend.
This week we start with a discussion of logistic regression and our first encounter with classification problems. This serves also as a motivation for introducing gradient methods since we no longer end up with nice analytical expressions for the optimal parameters beta.
The plans for this week are
Material for the active learning sessions on Tuesday and Wednesday.
Lecture from last week on the bias-variance tradeoff
Resampling techniques, cross-validation examples included here, see also the lectures from last week on the bootstrap method
Exercise for week 38, see also the whiteboard notes from week 37 at https:...
Dear all and welcome back to FYS-STK3155/4155.
We hope you all had a great weekend. The plans for this week focus on
1) The lab sessions
a) brief reminder from last week on expectation values. The exercises week focus on this topics and can all be reused in project 1. Take also a look at Wessel van Wieringen's article at https://arxiv.org/abs/1509.09169. This is a good read if you are somewhat rusty on expectation values and more.
b) Else, we will focus on work on project 1
c) A small discussion on scaling with examples. A jupyter-notebook will be sent separately. This will be discussed during the first hour of each session.
2) Lecture on Thursday September 7
a) we will focus on a statistical interpretation of Ridge and Lasso regression...
Dear all,
the files for project 1 should now be available. We have set a tentative deadline to October 9. See the folder https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects/2023/Project1Links to an external site. You can find the pdf file and latex file in the subfolder PDF. We will discuss the project during the coming lab sessions. Note also that the weekly exercises align with the project, that is you can reuse the exercises in the project.
Best wishes to you all.
Dear all, here are the plans for the coming week:
- Material for the active learning sessions on Tuesday and Wednesday
- Summary from last week on discussion of SVD, Ridge and Lasso linear regression, first 30-40 mins of each session
- Recommended Reading: Hastie et al chapter 3, see https://link.springer.com/book/10.1007/978-0-387-84858-7
- Presentation and discussion of first project and exercises for week 36
- Material for the lecture on Thursday September 7
- Linear Regression and links with Statistics, Resampling methods
- Recommended Reading: Goodfellow et al chapter 3 (till 3.11) on probability theory, see https://www.deeplearningbook.org/
- See also Murphy, sections 2.4 (Gaussian distr...
Dear all, first of all we hope you had an excellent weekend. we look much forward to welcome you back to this week's exercises and lectures.
This week we will discuss and work on the exercises for week 35 (all relevant for the start of project 1 next week). The material needed for these exercises is covered by the first part of the weekly slides for week 35, see the material before the heading Material for lecture Thursday, August 31. This refers to the material till slide 41 in for example the https://compphysics.github.io/MachineLearning/doc/pub/week35/html/._week35-bs041.html
The slides 1-41 contain also several examples and derivations relevant for solving the three exercises we will work on this week during the lab sessions.
Please do read this material before coming to the various lab sessions. Also, please do send us your questions before the sessions start. We will spend 15-30...
Overview of first week
First of all a warm welcome to you all.
The sessions on Tuesdays and Wednesdays last four hours for each group (four in total) and will include lectures in a flipped mode (promoting active learning) and work on exercices and projects. The sessions will begin with lectures and questions and answers about the material to be covered every week. There are four groups, Tuesdays 815am-12pm and 1215pm-4pm and Wednesdays 815am-12pm and 1215pm-4pm. Please sign up as soon as possible for one of the groups. Max capacity per group is 30-40 participants.
On Thursdays we have a regular lecture. These lectures start at 1215pm and end at 2pm. The first week we wtart with simple linear regression, a repetition of linear algebra and elements of statistics needed for the course.
- August 22: Presentation of the course, aims and content. Introduction to software and re...