Entrance hall of the ICMSLearning graphical models in high dimensional settings

In April 2017, ICMS hosted a workshop on Learning graphical models in high dimensional settings

The purpose of this workshop was to help meet the theoretical, methodological and applied challenges in extending graphical models to higher dimensional settings. It has two major objectives:

  1. To promote the development of new cutting-edge research at the interface of theory, modelling and applications for multivariate dependences, in order to face the challenges presented when trying to model large and complex data sets. 

  2. To stimulate the collaboration between researchers from different institutions, in particular to provide opportunities for young researchers to establish new collaborations and partnerships.

This workshop was very busy with 26 presentations, a poster session and a sold-out public lecture by Peter Bulhmann, (ETH Zurich), Cause or effect? Looking beyond correlation.  The public lecture included a discussion of the correlation  between a country's consumption of chocolate and award of nobel prizes.  To encourage scientific progress, Peter distributed chocolate (toblerone, off course) to the audience.

 

Robin Evans, Oxford, (left) introducing the public lecture by Peter Buhlmanm, Zurich (right)


 

Whilst the workshop was on, we took the opportunity to speak the delegates in a bit more detail.

Sofia Massa, University of Oxford

 

Sofia Massa is a Senior Statistician in the Nuffield Department of Population Health, University of Oxford. She has an MSc in Mathematics (University of Udine, Italy) and a PhD in Statistical Science (University of Padua, Italy).   Her main research interests are on probabilistic graphical models on aspects related to evidence synthesis, gene-set analysis and neuropsychology modelling.

Tell me about today's event and your role in it

I am one of the organisers for the workshop and  I was involved in preparing the proposal and shaping the programme.

What brought you to this area of research?

I’ve been working  in graphical models research since my PhD studies. My  initial research was motivated by some open questions coming from biology and genetics . I then explored many other directions within this area.   Graphical  models are now used in many different contexts  and they represent a very active area of research motivated from both theoretical and applied open questions.  The current  high-dimensional scenario is presenting many interesting and challenging problems to be solved. 

Other than exploring maths, what are the benefits of taking part?

I think the main benefit is that you are able to talk with everyone.  It isn’t big or intimidating.  There has been lots of discussion and opportunity to meet new people.  There are people from different areas and different backgrounds and this brings fresh insights and new perspectives into the research topics.

What will you take back to your [day job/research/studies]?

Quite a few insights related to high dimensional settings, causality and design of experiments.  I’ve made connections with researchers who offer an alternative point of view because of their different backgrounds.  There has been a lot of applied talks at the workshop and it has been interesting to hear about new challenges coming from more complex type of data available.  

Have you met interesting people, and if so, what connections have you made?

I think we’ve managed to generate a sense of belonging to a wider community of researchers working on related problems.  There is a growing network of researchers in this area in the UK and beyond.  It is good to see that geography is not a barrier for developing this network.

Do you have any advice for first-time ICMS attendees?

It is my first time at ICMS.  I would suggest to make the most of what is a great opportunity.  Get involved 100%, speak to people and be part of the debate.  This is especially important for Early Career Researchers.

Have you been to many other conferences? How does ICMS differ?

The main difference has been the lack of parallel sessions.  This has been great as you are able to follow all the talks and hear everything that is going on.  Also as most of the delegates are staying in the same place, there is more a feeling of being hosted at this workshop.

Do you have any thoughts regarding how we can raise the profile of maths?

Make it accessible.  Having public lectures where people can see the link between the theory and applications can help.  It is important that we convey  the importance of research at all levels, (e.g. schools and wider society). 

Do you have any thought on how diversity in mathematics can be improved?

I believe the UK is trying to change.  Recent initiatives such at the Royal Society unconscious bias training, University Athena Swan initiatives shows there is a commitment to changing things.  There is still, however a lot to do.  Role models are important, especially for students and researchers at all levels.

 

Soren Hojsgarrd, University of Aalborg, Denmark

Søren is Head of Department of Mathematical Sciences and Associate Professor of statistics at Aalborg University. Aalborg is situated in the North West of Denmark.

Tell me about today's event and your role in it

I was invited to the workshop to sit and listen and enjoy the talks.  This, I have done.

What brought you to this area of research?

These days I am a bureaucrat, Head of Department.  Before that, my area of research was graphical models.  I fear I am slightly rusty these days.  It is good to meet people in this area again and find out what is going on.  It is important to keep up to date, it is good for the brain to be re-challenged.

Other than exploring maths, what are the benefits of taking part?

They are a great bunch of people in this area, and the food has been great!

What will you take back to your [day job/research/studies]? 

Unfortunately, when I get back my priority is going to be to catch up on the extra work I’ve missed because I’ve been away from the office.

Have you met interesting people, and if so, what connections have you made?

There are quite a few new faces, but it is too early to tell where this will lead.

Do you have any advice for first-time ICMS attendees?

It is my first time at ICMS, but no advice.

Have you been to many other conferences? How does ICMS differ?

I think small workshops, in general work well.  They tend to be more interactive and you have the chance to talk with all/most of the participants.

Do you have any thoughts regarding how we can raise the profile of maths?

Yes, I think the UK is better in this area than Denmark.  There is a British tradition in maths/stats of being more diverse, and outgoing.  It is a tricky thing to do, maintaining the quality whilst widening the appeal. 

Do you have any thought on how diversity in mathematics can be improved?

I think role models play a vital role.  There is growing concern about access for people from disadvantaged social/economic backgrounds.  I don’t have an answer but I do think it is a concern.

 

 

Return to ICMS Blog postings list.