Tommy

A rainy and dull day today, such terrible weather for the end of May. I had a wet walk with Xena that turned in to a very muddy walk and she needed a good wash afterwards.

I decided to take a photo of Tommy today as today's entry is all about him. After graduating with a masters in engineering (machine learning) last year, he decided to take a year off to explore his career options for further research (PhD programmes) in his area of interest. He spent the next few months doing lots of reading and maths, and eventually he managed to write a scientific paper about a research idea that he had, and submitted it to a symposium (mini-conference) for his research area. Over the next few weeks he received some favourable reviews of the paper from the acceptance panel, and then last night he heard that his paper has been accepted at the symposium. It is quite a big deal to have a first author (meaning you are the primary researcher behind a project) paper published before one starts a PhD program, and a very big deal to have a single author paper (meaning you did not have the help of an academic supervisor or research team) at this point in one's career. The title of his paper is 'Structured Partial Stochasticity in Bayesian Neural Networks'. His area of interest has always been Bayesian Neural Networks which put very basically is to do with making machine learning models (neural networks) better at knowing what they do not know*. He is currently doing a 6 month internship where he gets to use his knowledge of Bayesian machine learning so it's all very relevant for him.

He will be going to the International Conference of Machine Learning in Vienna in July to present his paper at the Advances in Approximate Bayesian Inference symposium.

Sick of all the politicans' promises?  Bridgerton is the perfect antidote to watch in the run up to our general election in July. Ridiculous over the top outfits, beautiful people, frivolous society gatherings and romance - so much better than hearing some of the ridiculous things that the next wannabe-leaders are pushing in their campaigns.

* Tommy once explained it to me very simply, in the case of programming driverless cars, its getting the system in the car to recognise that the obstacle on the side of the road is not a bollard as would be a normal object on the route, but rather a small child wandering too close to the edge of the road.