Hello! My name is Alex Murphy. I’m interested in the computational processes that underly language representations in man and machine. I first achieved my Bachelor’s in Linguistics at the University of Bangor, Wales in 2013. I then spent two years studying Language Technology on a joint programme between The University of Reykjavík and The University of Iceland. I then moved to the University of Copenhagen in 2015, Denmark, where I spent two years studying IT & Cognition. Since 2017 I have been a PhD student at the Computational Cognitive Neuroimaging Group at the University of Birmingham. In 2018 I spent 4 months as a Technical Intern at Google Brain, London (Language Group).
I am currently researching how we can (a) use the latest machine learning techniques to “read the brain” and learn the neural characteristics that are associated with human language processing and (b) try to incorporate human language processing data into more traditional natural language processing (NLP) pipelines. I work primarily with electroencephalograpgy (EEG) but also have a deep understanding of functional magnetic resonance imaging (fMRI). I was the sole Teaching Assistant (TA) for the Master’s level course, Advanced Brain Imaging in Cognitive Neuroscience (2020 Spring Term, University of Birmingham), which focused on the acquisition, preprocessing and statistical analysis of fMRI data. During my time in Copenhagen, I was also an intern at the Danish Research Centre for Magnetic Resonance (DRCMR) for six months, where my primary role was writing analysis and experiment code to automate processing of fMRI data. I also spent a few months as a Teaching Assistant for the course, Discrete Mathematics (2017, IT University of Copenhagen).
Natural Language Processing
My primary interest is in part-of-speech tagging and parsing algorithms and how these actions are carried out in humans and computers and how the interaction of both research domains can be mutually beneficial. Human language data can benefit NLP because it adds another dimension to the data beyond what we currently have (this is analogous to going from black-and-white images to colour-images). NLP can also serve as a strong basis to analyse computational theories of human language processin, for example by studying the alignment properties between vector-space models and high-dimensional human neural recordings.
Brain-Computer Interface / Neurotechnology
I am very interested in the application of advanced machine learning methods to more accessible brain-reading technology such as the dry-electrode setups often used in the Brain-Computer Interface (BCI) world and how this technology might serve as the foundation for many extremely cool neurotechnological applications in the future. This is a hobby of mine and I hope my blog articles on this topic will be useful to other newcomers to the field.
Probabilistic Approaches to AI
I am an enthusiastic student of probabilistic methods applied to AI & Machine Learning. I was lucky to be part of the first ProbAI summer school cohort in Trondheim, 2018 and will hopefully be returning as a (virtual) TA for the 2021 edition.
I am a huge fan of aviation and currently studying for my pilot’s license. I like learning languages and meeting people of different backgrounds, sharing the human experience.
You can reach me by e-mail, Twitter, LinkedIn by clicking the icons below.