NZ RSE hackathon: Using machine learning to investigate simulation data
NeSI and the NZ Centre for Earthquake Resilience (QuakeCoRE) are partnering to offer a machine learning hackathon as a precursor to the NZ Research Software Engineering Conference (9-11 September 2020).
Over four days you will apply machine learning techniques to a real data science problem that comes from the NZ research community. This self-paced challenge is ideal for individuals or teams wanting to gain confidence in using machine learning to analyse real datasets. Researchers and research software engineers from all domains are encouraged to participate.
QuakeCoRE conducts large-scale, computationally intensive numerical ground motion simulations to assess earthquake risks around New Zealand. A particular challenge is that modelling low frequency ground motion can take between a few compute core-hours to thousands of core-hours on NeSI’s platforms.
In this hackathon, participants will develop a model to predict the number of core hours required to run a model, based on a number of features (problem size, number of of parallel processes, domain decomposition, etc). The solutions gathered will be judged based on the solution performance (details on judgement criteria TBD).
How to participate:
Start: 9:00am Friday, September 4th
End: 10:00pm Tuesday September 8th
Format / Location: Online. At 9:00am on September 4th the challenge will kick-off with all hackathon participants receiving a full challenge brief via email. For support during the hackathon we will have zoom 'office hours' where you can get help or guidance as required.
This challenge is designed so that a decent solution can be generated in under 10 hours. The problem size is suitable to run a on a laptop, you do not need to be a NeSI user to participate
The top solution for the hackathon will be announced on 11 September at the NZ RSE Conference.
Who is this event for?
This event is for researchers and research software engineers (RSEs) who are interested in machine learning and want to gain experience working on a real data science problem. Both individuals and teams are welcome to register.
If you are unsure if this event is for you please message firstname.lastname@example.org and we will be glad to guide you.
If you have any questions or would like more information about this event, please email email@example.com.