Andreas Bueff(bÜff)

Welcome to my site.

This site is about me, machine learning, and maybe some other stuff.

I am currently a PhD student at the University of Edinburgh, working with Dr. Vaishak Belle at the Artificial Intelligence and its Applications Institute.

During my time as PhD student, I worked as a research assistant with Nationwide and Edinburgh Business School to help stress scenario test their credit scoring models and investigate the use of counterfactuals for understanding model performance.

I then worked as an academic advisor with the NatWest Group to develop course material for an explainable AI course designed for data scientist employees. The course material was then expanded further in May of 2021 for the purposes of being taught as a graduate-level course through the University of Edinburgh's, Bayes Centre.

Prior to beginning my PhD in November of 2018, I pursued a Masters of Research with the Centre of Intelligent Systems and their Applications (now renamed the Artificial Intelligence and its Applications Institute). My supervisor was Dr. Vaishak Belle.

You can email me at andreas [dot] bueff [at] ed [dot] ac [dot] uk

Visit my linkedin profile, it still needs to be updated

You can check out my github however most of my repos are hidden at the moment.

My general areas of research in Machine Learning are in Explainable Artificial Intelligence, Reinforcement Learning, and Tractable Modeling. In the past, I have relied on Sum-Product Networks for my research into tractable modeling, and recently I have been exploring Inductive Logic Programming to explain Reinforcement Learning policies.

Some of my recent research has looked at differentiable Inductive Logic Programming (dILP) for the purposes of deriving non-linear explanations on data generated from mathematical formulas. In tandem with this, I have been exploring the use of dILP and relational reinforcement learning to derive similar non-linear policies to explain the state dynamics of RL environments.

Below you can have a look at my short-form curriculum vitae. My full CV is available as PDF.

• 11/2018 - present PhD Candidate, University of Edinburgh, School of Informatics, Artificial Intelligence and its Applications Institute.

  • 05/2021 - 07/2021 Teaching Assistant in Explainable AI, School of Informatics, University of Edinburgh, Expanded and recorded lecture content, developed assignments. Supervisor: Vaishak Belle

  • 11/2020 - 12/2020 Academic Advisor, NatWest Group and Edinburgh School of Informatics, University of Edinburgh, Developed lecture content. Supervisors: Vaishak Belle, Peter Gostev.

  • 07/2019 - 12/2019 Research Assistant, Nationwide and Edinburgh Business School, University of Edinburgh, Research on credit risk stress testing and counterfactuals. Supervisors: Vaishak Belle, Raffaella Calabrese.

• 09/2017 - 08/2018 MScR Program, University of Edinburgh, School of Informatics, Centre for Intelligent Systems and their Applications. (Distinction)

• 09/2016 - 08/2017 MSc Program, University of Edinburgh, School of Informatics, MSc in Artificial Intelligence. (Merit)

• 01/2014 - 12/2015 BSc Program, California State University of Sacramento, College of Engineering and Computer Science, BSc Program Computer Science. GPA: 3.63 / 4.0 (Cum Laude)

• 01/2012 - 12/2013 ASt, Sierra College, Business and Technology, ASt in Computer Science. GPA 3.65/4.0

• Andreas Bueff, Stefanie Speichert, Vaishak Belle; Probabilistic Tractable Models in Mixed Discrete-Continuous Domains. Data Intelligence 2021; 3 (2): 228–260. doi: https://doi.org/10.1162/dint_a_00064 [PDF]

• Andreas Bueff, Stefanie Speichert, and Vaishak Belle. Tractable querying and learning in hybrid domains via sum-product networks. In KR Workshop on Hybrid Reasoning and Learning, 2018. [PDF]

• Ghazan Khan, Andreas Bueff, Ivan Mihov, Nati Tessema, Javier Garrido, Chris Russel, Arash Parnia; Development of Transportation Asset Management and Data Collection System (TAMS) Using Mobile Applications. Procedia Engineering; Volume 161, 2016, Pages 1180-1186, ISSN 1877-7058 [PDF]