N.B. For an up-to-date CV summary, please see my about page

Hello and welcome to my new blog / personal website! My name is Giles, and I’m a researcher of particle physics and machine learning. At time of writing, I am on the cusp of defending my PhD in physics at the Instituto Superior Técnico in Lisbon, where I have spent the past 4.5 years living.

Currently I am working as a researcher at the University of Padova, having recently moved to Italy during a lull in the pandemic. My day to day work takes place mostly within the context of the CMS experiment at CERN, and my particular research interests revolve around developing, adapting, and applying machine learning methods to the various domain-specific problems we face in analysing high-energy physics (HEP) data.

During my PhD, I was an early-stage researcher within the AMVA4NewPhysics ITN, where I began my journey into machine learning (and previously blogged). My research mostly focussed on regression and classification tasks towards the search for the process of a pair of Higgs bosons decaying to a pair of bottom-quarks and a pair of tau-leptons. Such a rare process requires state-of-the-art sensitivity, in order to eek as much information as possible from the available data. To this end, I employed a variety of recent methods for training and applying neural networks (paper here), which in turn led to me developing a Python package to allow other researchers to also apply such methods to their own analyses of HEP data (LUMIN).

Prior to my PhD, I completed a 4-year integrated master’s in theoretical physics at Durham University; in which I wrote a thesis on the problem of classifying the hadronic decays of top-quarks using various algorithms. This was then followed by a 2-year postgraduate master’s by research in physics at the University of Glasgow; where I studied the mismodelling by Monte Carlo generators of the process in which a gluon splits to form a pair of bottom quarks.

My current research now also includes developing regression models for muon energy measurement in calorimeters (preprint), systematics-aware training methods for models (see e.g. Inferno by fellow AMVA4NP ESR Pablo de Castro and network coordinator Tommaso Dorigo), and end-to-end optimisation of detector design as part of the newly-formed MODE collaboration.

Beyond academia, I enjoy and noodling about on guitar and bass, cooking, and appreciating whiskies, wines, and ales. I’m also working towards my 2nd grade blackbelt in Shorinji Kempo.

I aim for this new site to function as both a continuation of my old blog, and a running CV of publications and presentations, etc. With regards to blog topics, I imagine I’ll be covering a mixture of technical explanations of physics and machine learning, academic life, travel, and maybe the occasional recipe or two. I make no promises about frequency or regularity, but I’m also on Twitter if you fancy following me there.