Sometime in April I stumbled upon the website of Oxford Machine Learning Summer School (OxML for short) and their call for applications. They proposed lectures in 2 tracks: ML in Health and ML in Finance. Wanting to explore some new research directions, I didn’t hesitate even for a minute.

 

In this short series of blog posts I want to introduce you to my thoughts about the school, how to apply, what are the costs and what did the programme look like this year. It’s going to have 3 parts:

1. General information, application, costs (this one)
2. My thoughts on MLxHealth track (TBA)
3. My thoughts on MLxFinance track (TBA)

 

Note that I participated in the 2022 edition and the information may change over the years. 🙂

 

General information about the Oxford Machine Learning Summer School

 

The school is organised by AI for Global Goals with a partnership of CIFAR and Oxford’s Deep Medicine Programme. It’s taking place at the University of Oxford, but you can also participate in the school online.

 

This year it consists of 3 parts:

 

1. Fundamentals track – 3 days of online lectures with introductory information about ML for everyone
2. Health track – 4 days of in-person and online lectures about application of ML in health sector
3. Finance track – 4 days of in-person and online lectures about application of ML in finance sector

 

Application to OxML

 

Application process is done through Google Forms. It starts in February and lasts till April. 

 

Apart from the standard information (name, surname, etc.), you need to fill the following information:

 

1. What is your completed/awarded degree (choices: BSc, MSc, PhD)
2. What is your current affiliation (university or company)

 

You also need to rate your ML programming skills – in each score they explain what they mean, eg. my choice was “Very good (completed multiple ML research projects; experienced in multiple ML/DL libraries; I code almost daily)”. There was a higher score that would indicate that you basically can program anything ML-related and I was tempted to check it, but my scientific soul wouldn’t let me claim something so broad like anything. 😂

 

The last big question was to explain a significant ML-related project that I have done. I’m copy pasting my answer below.

 

One of the most significant projects I did was the paper I published for Interspeech 2020 as a first author regarding open set audio classification. It was immensely challenging from the research and time-frame point of view, but also because I had to manage the work of 3 other researchers who participated in the project.

In the paper, I described my proposed solution to the problem of acoustic scene classification in an open set setup – meaning when we have the possibility to either classify within known classes or reject the classification, which is largely required as the industrial requirement for ML classification products. My tasks included research and understanding of the previously proposed method, implementation, training and benchmarking of the new proposed solution, preparing the final paper as well as further presenting it during the Interspeech 2020 virtual event (oral). 

 

You also had to indicate if you want to participate in only 1 or both tracks and if you want to apply for in-person or online events. There were less places for in-person experience, so you could also state that you prefer in person, but in case you don’t get in – you still want to participate online.

 

I assume the decision was made largely based on your resume, because we also had to attach one. I think I attached this one, or at least I hope so.

 

Last but not least, you had to post links to your social media and online profiles if you had one (it was optional). I posted mine:

• LinkedIn
• Twitter
• GitHub
• Google Scholar

 

Costs

 

The first major cost is a cost of participation in Oxford Machine Learning Summer School. The organisers prepared a convenient table indicating what are the costs differences between virtual and in-person experience, as well as based on your level of experience and affiliation. 

 

	Virtual	In-person
Full-time students	80 GBP	350 GBP
Full-time academics	150 GBP	600 GBP
Other	300 GBP	900 GBP

 

 

Note that those are the costs of Fundamentals + 1 track, so if you wanted to participate in both Finance and Health tracks, you had to double the amount. I paid 700 GBP in total for that (full-time student, in-person). This also includes a really, I mean really, good lunches and coffee breaks. Each day we have the experience of dining in one of the halls and we get a 3-course meal. You could also indicate any dietary requirements that you had (I need dairy free and paid nothing for it additionally).

 

Next, there are costs of living in Oxford. There was accommodation available in the college where the event takes place (St. Catherine’s) and I decided to go for it. It cost me 930 GBP for 10 days for a single room with my own bathroom. The place is very clean and nice and a really nutritious breakfast is included in the price.

 

As you can see now, as we only have breakfast and lunch covered, you need some additional money for potential dinner or if you simply want to go out for a beer with friends. I would assume 10-15 GBP per day for such expenses.

 

The travel costs highly vary depending on where you are from. I travelled from Warsaw to London and then took a bus to Oxford. It cost me around 200 GBP for both flights and 30 GBP for both bus tickets. The cost was high because I booked it really late – I wasn’t sure about going – but you get the information about the form of attendance in May so you still have a lot of time till August to book cheaper flights.

 

To sum it up:

• Registration fees (including dairy free lunch & coffee breaks): 700 GBP
• Accomodation (single room with own bathroom + breakfast): 930 GBP
• Travel expenses: 230 GBP
• Other costs: let’s say around 100 GBP

 

Which means I roughly spent 2k GBP on a 10 day trip to Oxford from Warsaw, Poland. 

 

Final thoughts

 

In the next 2 parts I will describe the experience of participating in both tracks and the conference itself (spoiler alert: it’s awesome!).

 

If you have any more questions about OxML – don’t hesitate to contact me via my social media. Take care!