Tools for AI/ML research in life sciences is an event series by the SciLifeLab Data Centre aimed at life science researchers who use machine learning methods in their work. The goal of the events in this seminar series is to provide introductions to different tools for ML research but also to foster discussions around our practices and how they can be improved. The events takes place virtually (over Zoom) and are open to researchers in Sweden and beyond. Each event is scheduled for 60 minutes, consisting of a talk and an extended discussion.

Contact: serve@scilifelab.se
Scientific lead: Prof. Ola Spjuth, SciLifeLab Data Centre and Uppsala University

Upcoming events

---

Past events

Apr 18, 10:00-11:00, Zoom webinar

Leveraging supercomputers for data-driven life science research: examples from Berzelius

Soumi Chaki and Xuan Gu (National Supercomputer Centre, Linköping University), Kimmo Kartasalo (Karolinska Institute), Claudio Mirabello (Linköping University)

Abstract: Powerful GPU supercomputers allow to carry out large-scale experiments in a fraction of time that it would otherwise take as well as pursue research projects that would otherwise not be possible. Life science researchers at Swedish universities can apply and make use of the Berzelius AI/ML cluster at the National Supercomputer Center in Linköping. In this one hour long webinar two researchers will present how they use Berzelius in their work and what research it allows them to carry out. In addition, we will present the technical specifications of Berzelius, the resource allocation procedure, as well as the available user support. We will give a short demonstration on how to request GPU resources on Berzelius and run scripts/software on the allocated GPU(s). This webinar is particularly tailored for researchers interested in and working with AI/ML methods.

Feb 2, 10:00-11:00, Zoom webinar

Practical intro to GPU programming in Python and Julia

Yonglei Wang, PhD, Research Software Engineer and HPC application expert, ENCCS

Abstract: Availability of Graphics Processing Units (GPUs) has transformed the way we work with machine learning and data science challenges in life sciences. The parallel processing capabilities of GPUs have allowed training of ever more complex models, allowing researchers to analyze large biological datasets with unprecedented efficiency. However, in order to make use of the potential that GPUs offer we need be able to write fitting machine learning model code and analysis pipelines. In this webinar ENCCS will present some practical tips about what to keep in mind and how to optimize your code when running analyses on GPU hardware. This webinar will be most useful to researchers who already work with large datasets and would like to improve their understanding of how to work with GPUs. At the end, the participants will also be given an overview of online materials and in-person courses where researchers can learn about this topic in depth.

Nov 16, 13:30-15:30, SciLifeLab Solna campus

Building and sharing machine learning demo applications within life sciences: a practical tutorial

AI Data Engineers from the SciLifeLab Serve team

Abstract: It is becoming increasingly popular to share machine learning models with the community as web applications with an easy-to-use interface. Users can then adjust parameters or submit their own input and see the predictions generated by the underlying model. This tutorial is aimed at researchers working within life sciences who work with machine learning models but do not have the skills to build applications for web. During the tutorial we will start from a trained model and demonstrate step by step how you can create a graphical user interface for your application, prepare it for deployment, and make it available on the web with a URL. We will demonstrate the use of specific tools which make this process easy and doable in under an hour. The workshop will last for 2 hours with a break in the middle. We have room to accept 30 participants, on the first come first served basis. Those registered after that will be placed on the waiting list.

Oct 19, 10:00-11:00, Zoom webinar

Using containers to simplify ML training on Berzelius and other supercomputers: beginner-friendly introduction

AI engineers from the SciLifeLab Data Centre and application experts from Berzelius (NSC, LiU).

Abstract: A common approach to train machine learning models is to first create a prototype using a small dataset on a local machine to verify that it works and thereafter use a large scale compute infrastructure such as Berzelius for the full-scale training. One of the challenges with this approach however is incompatible systems in terms of differences in available software packages, versions, etc. An effective way to solve this issue is to use a container solution. Using a container environment allows a highly portable workflow and reproducible results between systems as diverse as a laptop, Berzelius or EuroHPC resources such as LUMI for instance. During this beginner-friendly event, we will introduce and demonstrate how to work with containers on Berzelius (Apptainer and Enroot) using an example from life sciences, starting from raw data and finishing with a trained model. During the Q&A session, the Berzelius life science support team will answer your questions.

Sep 08, 10:00-11:00, Zoom webinar

Modelling life science data in big pharma, academia, and startups – Differences, Examples, and Some Learnings

Andreas Bender
Professor of Molecular Informatics, University of Cambridge

Abstract: Life science data, modelling methods, and organizational environments in which they are implemented vary widely – from big pharma, to academia, and start-up environments (and beyond). This presentation will partly cover science and modelling, but there will be an additional focus on the differences between the above environments (all of which the presenter has worked in), and how they influence the approaches, as well as the eventual result and product to be delivered.