COMP9417 17s1 Course Noticeboard Machine Learning and Data Mining
Some answers for final exam questions (Posted by Mike Bain, Tue 20 Jun 6:32pm)  

See under "Lectures"

Final Exam Sample Questions (Posted by Mike Bain, Mon 19 Jun 2:30pm)  

See under "Lectures"

Revision Guide plus Tutorial Answers (Posted by Mike Bain, Mon 19 Jun 10:25am)  

See under "Lectures" for a guide to revision for lectures based on learning objectives.

Also, see under "Tutorials" for answers to last two tutorials.

Assignment 2: submission (Posted by Mike Bain, Fri 2 Jun 11:30pm)  

Please note the following updates to the procedure for submitting assignment 2:

1) submission instructions have been added to the spec
2) each *team* should make *one* submission
3) ensure *all* team members names *and* student numbers appear on the 
        front page of your report!
4) the limit on submission size has now been increased to 2MB
5) if your submission will exceed 2MB, consider if you really need to 
        submit all of that. If you conclude that you do, then submit your 
        *report* via give, and in it provide a download link to the 
        additional material.

Schedule for Week 13 (Posted by Mike Bain, Mon 29 May 3:38pm, updated Tue 30 May 9:54am)  

This week we have the following schedule:

Tutorials: usual times and places, material from last week’s lecture on "Probabilistic Graphical Models" by Edwin Bonilla

Lecture: Tuesday 30th May, 6-9pm CLB 6

(1) Guest Lecture: "The commercial realities of machine learning” Andrew Botros, Phd. Commonwealth Bank

This lecture will cover topics such as organisational attitudes to machine learning, differences in analytical cultures, machine learning processes in production, dealing with biases in commercial data, managing hype and managing change.

Dr Andrew Botros is Lead Statistician at the Commonwealth Bank and formerly Principal Engineer at Cochlear Limited, where he created the cochlear implant industry’s first machine-learned system.

Here are a couple of resources related to the guest lecture: Breiman (2001), and IBM (2015)

Additionally, some of the guest speaker's work on machine learning applied to modelling musical instruments, and automating the calibration of cochlear implants.

(2) Review of course topics and Q&A on exam preparation: following the guest lecture I will stay around up to 9pm if needed to review the main topics we covered and answer questions on the final exam. If you can’t make this time, I will be available for consultation on Wednesday 31st May 2-3pm. If you can’t make that either contact me and we will try and arrange some further time for consultation.

Note: a number of people have asked me this - the content from the lecture in Week 12 on "Probabilistic Graphical Models" by Edwin Bonilla will NOT be in the final exam.

Assignment 2 (Posted by Mike Bain, Mon 24 Apr 10:39pm)  

Refer to specification under "Assignments". More topics may become available. If you have a proposal for a topic then send it for consideration - see spec for more details.

No lecture or tutorials on Tuesday 25th April (Posted by Mike Bain, Mon 24 Apr 8:59pm)  

Reminder: due to a public holiday, there will be no lecture or tutorials on Tuesday 25th April.

Mid-semester exam (Posted by Mike Bain, Mon 10 Apr 3:53pm)  

DATE: Tuesday 11 April

TIME: 6:15-7:15 pm (duration is 1 hour)


What will be examinable:
Lecture topics covered in the course, starting from the lecture on"Supervised Learning - Regression" up to and including where we concludedthe lecture of Week 5 (Tuesday, 28 March) which was on slide 61 of thelecture on "Tree Learning".

What you should bring to the exam:
UNSW-approved calculators only.

Sample questions are under the "Lectures" link for Week 7.


Assignment 1: Spec released, deadline extended (Posted by Mike Bain, Fri 24 Mar 9:44am)  

See the link under "Assignments"

Reminder on using course forum (Posted by Mike Bain, Mon 13 Mar 8:42am)  

We are using Ed for course discussion this semester.

Unless your question is a personal issue, please do not send us direct email.

Instead, post your questions on Ed so that others can benefit from the answers.

We encourage all students to regularly check the course discussion page, and also answer and comment on any questions or posts that interest you.

Tutorial for Week 3 (Posted by Mike Bain, Fri 10 Mar 11:07pm)  

This week the tutorial is on Regression and uses the Jupyter notebooks which requires Python. There are a number of ways to get started. Here is one using the command-line:

  1. Install Jupyter
  2. Download the "Lab1_Questions.ipynb" file from the course website under the link for Week 3 tutorial (you may have to rename the file to remove the suffix .txt)
  3. Change to the directory where the file is located
  4. At the command-line type "jupyter notebook"

A browser window should appear with the notebook open in it.

If you don't want to install Jupyter, you can use the online "Try Jupyter" approach:

  1. Download the "Lab1_Questions.ipynb" file from the course website under the link for Week 3 tutorial (you may have to rename the file to remove the suffix .txt)
  2. Go to Try Jupyter
  3. Click on 'Upload', navigate to the correct directory and select the "Lab1_Questions.ipynb"
  4. Click on the new 'Upload' button, then select the notebook
Please use forum on Ed (Posted by Mike Bain, Fri 10 Mar 10:45pm)  

We will be using Ed to host the course forum. By now everyone should now have received an invitation by email. Please ensure you respond to this email and enrol yourself, ask questions and join the discussion!