If you are using the sentence from lecture slides, tutorial materials or assignments. We will not allege plagiarism.
If you are copying sentences in large portions (i.e. >= 65%) from online resources such as Academic Papers, GitHub, Chegg, Towardsdatascience.com, StackOverflow etc. It will be recognized as plagiarism.
If you believe the submitted answers are your own work, you don't need to worry about that. And the previous notice is only for some students not everyone in this course.
Sorry for the confusion.
[It only applies to some students]
Dear all,
We've identified significant level of plagiarism for some students submitted final exam answers. It is a contravention of the UNSW exam rule. Although it is an open-booked exam, it does not mean that you can copy the text from an online resource or academic paper directly.
We would like to give you a chance to explain your motivation and behaviour and not escalate your case to UNSW Conduct and Integrity department for further assessment which will lead to 0 in this course.
Please send a statement to cs9727@cse.unsw.edu.au by tomorrow 10:00 am as we are not willing to delay the marking process.
If we do not receive your response at the given time period, we will escalate your case to UNSW Conduct and Integrity Department for central management.
Thanks,
Lina
Congratulations on finishing the final exam. If you have a network issue, please follow the guideline to submit your answers ASAP.
We know that you may not be able to finish all the questions in time and stressful with the exam hurdle. Don't worry about that, we will scale up if necessary. For now, just enjoy your short break and prepare for the upcoming T3.
For those students who are interested to conduct the thesis (for undergraduate) or research project (for postgraduate) or Master of Philosophy or PhD with me and my group. Please do not hesitate to drop me an email with your CV and your recent academic transcript, we're happy to arrange a chat with you if we find a good match.
There are some students who are confused about how to deal with the situation: if the rating matrix has multiple missing values, how to calculate the similarity. You can check the example here for user-based methods.
Moreover, we've released the code solution for the sample exam Question 4.2.
If you find any questions in your final exam tomorrow, please send an email to cs9727@cse.unsw.edu.au. If we find some typos or problems with the questions, we will send a notice within the Inspera. So you don't need to monitor the updates on WebCMS3 or your mailbox.
Dear all,
Just a reminder that the final exam time is from 1:45 pm to 5:00 pm on 24th Aug. You can access the exam paper on Inspera ( https://unsw.inspera.com/ ). Please ensure you will log in on time. Please find a place with a good internet connection; if you suffer from an internet issue, don't panic (please read the instruction about how to deal with the internet issue before the exam starts here ).
You will have eight questions (36 sub-questions in total) in your final exam. To avoid confusion, the sample exam will close on 12 pm, 24th Aug so that you will only have one active test in Inspera.
You may need to prepare several white papers and pens as you will need to do some hand-written derivations or calculations (you will need your phone or other devices with a camera to take a photo/scan of your works). Please ensure all your steps are clear and easily recognised in your submitted pdf file. If you are more confident using Word/LaTeX to produce the answer pdf file, you are free to use them.
You will have some coding questions in your exam. Please open your editor/IDE before the exam starts so that you can avoid spending time on IDE's initialisation. Also, please ensure that your compile environment contains numpy and pandas so that you can test your code.
You will ask to conduct some calculations using Inspera, so please ensure you are familiar with Inspera's math tools. For example, how to write a fraction. You should be able to find a practice question in your sample exam.
You may be unable to finish all the questions in the three-hour timeframe as they may be more challenging than the sample exam. Don't worry; we will make some adjustments if we find the majority of you cannot finish the exam in time.
Moreover, Inspera has built-in plagiarism detection modules. Please do not send your answers to your classmates. If plagiarism is found in your final exam, you will get a 0 in this course.
We will have one last offline consultation session on the upcoming Monday. If you have any questions, please come and ask.
Good luck with your final exam.
Lina
Dear all,
The assignment 2 marks are released. The bonus mark is automatically added to your assn1 or assn2 marks (based on the private leaderboard which was released on Monday). Some of you may not be able to see your assignment 2 mark because of plagiarism or no submission. If you believe there is a problem with your mark, please email us at
cs9727@cse.unsw.edu.au
.
The exam consultation times are released, please refer here .
Again, please ensure you are confident to use the Inspera before the final exam. We will send out another notice about the final exam when the date is approaching.
Lastly, good luck with your final exam.
Lina
Just a quick reminder about the assn2 report. Do not forget to include your Kaggle Team Name (that one that shows on the leaderboard) in your report. Otherwise, you may lose marks for the leaderboard challenger and bonus.
Moreover, don't forget to select two final submissions which you would like to use for the private leaderboard!
Just kindly remind that the assignment will be due in 2 days, so far only 17 students (less than 50% of total students) have participated in the Kaggle competition.
For the sample exam, we have provided the pdf version paper for revision. However, please log in to the Inspera ( https://unsw.inspera.com/ ) to play around and ensure you know how to enter the answers, type the math formulas, download sample code, upload files etc. We will not provide the sample solution for the sample exam as you can most of them in your tutorial solutions.
We will have one exam consultation each for online and offline students to answer questions regarding the sample exam or final exam. Time will be announced later.
We are currently recruiting one research assistant. You are expected to conduct machine learning related works. The pay rate is ~ $68/hr + super. You have to be an Australia Citizen . If you are interested, please send me your CV and transcript.
In order to clarify the difference between the sequential recommendation and neural network-based recommendation, we've published detailed marking criteria for assignment 2 and you may refer to it for more information.
Based on the feedback that we received on yesterday's lecture, we've published the sample exam which you can find it on the corresponding section on the main menu.
We're receiving some feedback about assignment 2 performance marking. Hence, we decide to change the marking criteria. Now, the performance mark comes as a bonus. And the tasks come to be more specific to replace the original performance mark scheme.
For more information please refers to the new spec. Based on that, the deadline has been extended to 5th, August.
The goal of assignment 2 is to test your understanding of the recommendation algorithm. You are free to submit the hard-coded results to test the format. But, you will not receive any mark if you submit these hard-coded results as your final submission (for example, all the predicted values are hard-coded as 5). If your final submission is hard-coded, you will receive 0 for this assignment.
Your results should be generated by a model/algorithms. You are free to use any recommendation algorithms or machine learning algorithms (there are lots of online resources, you can take them as a reference, but do not copy them directly).
Assignment 2 is released.
You should have already received your assignment 1 feedback from COMP9727. Here are the basic statistics about the marking:
Mean: 16.91/25, std: 6.8. Distribution: 36.4% HD, 18.2%DN, 9.1%CR, 9.1%PS. 27.3% < PS (with 15% of no submission or empty file).
Dear all
Thanks for attending this week's guest lectures. I hope you find these useful and inspiring.
Lucas introduced the different perspectives and practices between academia and industry. He would like to add further explanations as academia and industry definitely do share commonalities and be mutually reinforced. Here are his remarks
"
The industry does require complicated algorithms!
Perhaps I didn't make it clear enough. It's not correct to say it doesn't require, just that *most of the time* we will realize that we don't need so complicated algorithms. But, I would say we need to find very concrete use cases of those 'complicated algorithms', and 'persuade' or 'educate' the directors/managers about why should we invest in those algorithms (e.g., you see improvement in offline metrics). And as mentioned in the lecture, it also depends on other stakeholders (e.g., the product team, the design team, etc.).
Regarding academic papers, it's definitely a need to push the boundaries of research. One of the examples is the attention mechanism. It appears to perform very well in the industry across different domains. However, it would be even better if the researchers can 'explain' why they design such algorithms. Because if we just make the model complicated without good explanation and good ablation studies/analyses, there's a high chance that we just 'fit the curve' to yield better accuracy, in my opinion. "
Lina
We've discussed with the school about the potential extension of the deadline. Finally, we are allowed to extend the deadline of assignment 1 to Sunday, 3rd, June 23:00:00 AEST . And no further extensions are allowed unless you are awarded special consideration by the university.
A few students are asking the Q5 parts b and c both in the forum and consultation. To make it clear we have updated the spec: move the for all operators in part b to make it clear and make it provable. For part c, those 3N points are chosen uniformly at random (Thanks to Lawrence and Rui), and now you should be able to finish part c with the hints provided. And part d should not be affected.
Wednesday’s lecturers will be delivered by guest lecturer. Due to company’s privacy policy, we can not provide the recording to the guest lecture Moreover, the contents covered in guest lecture will not be examined.
We will have Raiz Investment for the first hour and Apple for the second hour.
Hi all,
Just a friendly reminder that the census date is 26 Jun 2022, 11:59 pm. It is your last chance to drop T2 courses without financial liability as well as the academic penalty.
We've noticed some difficulties for calculating Bayes CF on Assignment 1 Q3 when doing the round-up. You may find that most scores are 0.00 if you conduct the round-up. Hence, we decide to remove the round-up requirement in the marking criteria. Please report the full results in your Q3 part b as well as code output. Sorry for the confusion.
Consultations will take place next Monday 2:30 pm - 3:30 pm. Offline one will take place in Consultation Room 402, K17, CSE building. The online meetings link will be published in the Consultations section. If you have any questions regards to the Assignment, tutorials or etc, you may come to ask.
Please check the assignment page regularly as we may post some changes about the spec such as typo fixing (we are maintaining a change log so that you can quicky find the changes). Moreover, before you post your question in the forum, please ensure it has not been asked previously.
Assignment 1 is released which will be due on 1st, July, 17:00 AEST. We encourage you to write your report in LaTeX and a sample file is provided. If you have any problems with the assignment please ask on the forum.
Important: Reproducing, publishing, posting, distributing or translating this assignment is an infringement of copyright and will be referred to UNSW Student Conduct and Integrity for action. Hence, the assignment spec is encrypted so that you will not be able to modify or copy the text inside.
The slide of week2 is updated for typo fixing. There will be no lecture next Monday due to the public holiday.
Good luck with your assignment and enjoy your long weekend.
School advised that all the online tutorials should be conducted via Teams. Hence, we have moved all the online tutorials to Teams. Please go to your Microsoft Teams to check whether you are added to your tutorial group. The zoom link will no longer be available. Moreover, we have adjusted the tutors' allocations for the online tutorial to ensure that all the tutorials will start on time.
The consultations will happen each Monday 1pm- 2pm from the week after next week (i.e., from Week 4). There will be two parallel sessions (online and offline). More details will be published in the consultation section in the main panel.
Dear all,
We are informed that there are still 17 students who are oversea and not able to enrol on the online tutorial. Hence, we decide to convert the W13A offline tutorial into an online one. If you would like to attend the face-to-face tutorial, please consider enrolling W11A or W12A.
For those students who are oversea, please enrol on W13A or W14A.
Sorry for the inconvenience.
Dear all,
Based on the current CSE policy regards to the COVID-19. You should wear a face mask when attending face-to-face lectures and tutorials if the physical distance can not be maintained. If you feel unwell, please do not come to campus and join the online classes if necessary.
Tutorial 1 is up, you can spend some time looking at it. In week 1's tutorial, we will do some revisions about the machine learning concepts and linear algebra etc.
Hope you have a good weekend and see you guys next Monday.
Lina
Hello everyone,
Firstly, welcome to the course COMP9727. I hope you look forward to learning about recommender systems. In this opening notice, I outline how the course is run in 2022T2. The course outline is up and you are welcome to check it through.
In this term, we are expected to have more face-to-face classes. As you may be aware, we all have all the lectures offline and most of the tutorials (3/4) offline. However, we may switch to entirely online if the COVID-19 restriction comes back.
We will mainly use WebCMS3( www.cse.unsw.edu.au/~cs9727 or https://webcms3.cse.unsw.edu.au/COMP9727/22T2 ) to manage the course (the very system you are reading this notice from). All announcements, lecture notes, tutorial notes, assignment spec, etc. are going to be posted here. Moodle is used to provide access to lecture recordings and exams.
Lectures
start from Week 1
:
We will deliver lectures on
Webster Theatre B
at the scheduled lecture time 16:00 – 18:00 Monday and 16:00 – 18:00 Wednesday (Sydney time). All lectures will be recorded.
Tutorials
start from Week 1
:
We will deliver all of the face-to-face tutorials on
Goldstein G07
at the scheduled time which you should know. We understand that some of you are still not able to return to Australia. No worries, we will have an online tutorial for you.
Forums
:
if you have any questions, you can post them in the forums. We will visit the forum regularly to answer questions.
Looking forward to meeting most of you soon next week.
Stay safe and enjoy your O-week.