Teaching‎ > ‎Zimowy 2017-2018‎ > ‎

Neural Networks Fall 17

This is the main website for the Fall 2017 course on Neural Networks and Deep Learning. Please come back for announcements.

Lectures: Tuesdays 8:30-10:00 room 119

Final exam: since there was no consensus on the exam date, we will have two sessions:
- Monday 5.2.2018 14:00
- Friday 9.2.2018 14:00

You can come to either of the exam sessions.

Make-up exam: Friday Feb 16 at 10:15 in room 202

Last days to hand in projects:
JCh: Fri 9.2 10-14
AL: Thu 8.2 10-14
Extension (only for pre-authorised students):  Friday Feb 16 at 12-14 in room 110 or 202.

Grade thresholds for the assignments are:
100 total (60 assignments + 40 project): 90: 5.0, 80: 4.5, 70: 4.0, 60: 3.5, 50: 3.0

Final exam grade thresholds (out of 75 + 10% from assignment):
70: 5.0, 62: 4.5, 53: 4.0, 44: 3.5, 35: 3.0

Assignments and lecture notes: https://github.com/janchorowski/nn_assignments

Course rules: course rules.pdf

Acknowledgments:
I thank Google for awarding the class GCE credits under the Google Cloud Platform Education Grants. 

RankingNN2017Fall


Lectures 13-15

posted Jan 30, 2018, 5:54 AM by Jan Chorowski   [ updated Jan 30, 2018, 5:54 AM ]

During the make-up lecture we have spoken about the SVM and about the PAC learning theory. The slides for PAC are here: https://github.com/janchorowski/nn_assignments/blob/nn17_fall/lecture_notebooks/14-PAC.pdf

The RL video that I wanted to play to you is here with the accompanying notebook.

Assignment 6

posted Jan 25, 2018, 3:11 AM by Jan Chorowski   [ updated Jan 25, 2018, 3:11 AM ]

I've posted assignment #6 which will bring the total exercise points to 60 and adds several bonus questions. Please get it from: https://github.com/janchorowski/nn_assignments/blob/nn17_fall/assignment6/Assignment6.ipynb.

Submission deadline:

  • Problem 1 is due last lab session of the semester before or on Wednesday, 31.01.18

  • Bonus problems are due Wednesday, 7.02.18 (you can show them to the instructor by prior appointment).


Make-up lectures

posted Jan 22, 2018, 2:22 AM by Jan Chorowski   [ updated Jan 22, 2018, 2:22 AM ]

We have selected to host a make-up lecture for Saturday, 27.01.2018. For those that can't make it, please watch lectures 7 & 8 from https://see.stanford.edu/Course/CS229 which cover similar topics.

Lecture 11

posted Jan 22, 2018, 2:19 AM by Jan Chorowski   [ updated Jan 22, 2018, 2:19 AM ]

Lecture 10

posted Jan 9, 2018, 9:53 AM by Jan Chorowski   [ updated Jan 9, 2018, 9:53 AM ]

Lecture 09

posted Jan 8, 2018, 5:01 AM by Jan Chorowski   [ updated Jan 8, 2018, 5:01 AM ]

We have talked about LSTM and RNN usage for real world problems.
LSTM demo is in the updated notebook from lecture 8.

The attached presentation is from the PLinML meeting, I have also shown it in class.

Lecture 08

posted Dec 12, 2017, 12:38 PM by Jan Chorowski   [ updated Dec 12, 2017, 12:38 PM ]

Lecture notebooks are here:

Projects
Please see the attached slides for information about the projects and reproducibility. If you want to do the reproducibility challenge, please select an article by putting you name into this spreadsheet.
Please prepare by the end of next week a 1 page description of the project topics (what is the topic, what resources do you have, what steps do you want to take). Please consult the topic with the instructors.

Assignment 5

posted Nov 30, 2017, 8:01 AM by Jan Chorowski   [ updated Nov 30, 2017, 8:01 AM ]

Is posted: https://github.com/janchorowski/nn_assignments/blob/nn17_fall/assignment5/Assignment5.ipynb

Since it took us more than predicted to prepare it, the deadline is:
Problems 1 and 2 last lab session before or on Wednesday, 13.12.17
Problems 3 and bonuses last lab session before or on Wednesday, 20.12.17

Please note that project descriptions will also be due 20.12.

Lecture notes 7 & update to HW 4

posted Nov 21, 2017, 5:18 AM by Jan Chorowski   [ updated Nov 21, 2017, 5:18 AM ]

The notebook with convolutions is here: https://github.com/janchorowski/nn_assignments/blob/nn17_fall/lecture_notebooks/07-convolutions.ipynb

I decided to exted the deadline for problem 7  (conv net) on HW4 until Wednesday 29.11.17. I have also added a bonus problem, also due  29.11.17.

Lecture notes 6

posted Nov 18, 2017, 4:36 AM by Jan Chorowski   [ updated Nov 18, 2017, 4:36 AM ]

The notebook is here: https://github.com/janchorowski/nn_assignments/blob/nn17_fall/lecture_notebooks/06-regularization-numpy.ipynb There's a bounty of 5 exercise points for changing the neural network implmentation to PyTorch.

The slides are attached.

1-10 of 16