Nowadays, artificial intelligence (AI) using machine learning (ML) is greatly influencing our life in numerous areas including medical science, energy, social science, and others. The key of such success applying machine learning to real world problems is that machine learning is able to learn to understand the patterns and phenomena inherent in observational data in a real-world scenario. Precisely, most of these applications could be formulated as some standard problems in ML such as prediction and inference. For example, utilising big data, ML models such as deep learning can successfully make image and speech recognitions, recommendations, and medical diagnosis.
In fact, the common task of these ML models is to learn the associations between the variables inherent in the data. After these associations are learned, ML models could be used to 1) make prediction and 2) infer the relationships between factor variables. For example, many ML models are substantially regression and classification models whereby the relationships between dependent and independent variables can be learned which can then be used for prediction and inference. However, the learned relationships between dependent and independent variables are only linear or non-linear associations; this relationship does not necessary infer the causal relationships between them. These causal relationships are the key points for interpreting ML models. Our objective of this project is thus to infer causal relations in ML models.
Typical example applications of causal inference include treatment effects of drugs and energy consumption effects; this project will focus on such applications. In more detail, this project will use observational data to estimate the causal effects. In some situations, randomized control trials (RCT) could be used, but these methods are usually expensive and biased as the samples are barely representative for the population. Precisely, this project targets devising reliable and scalable machine learning models for estimating causal effects using big observational data. To deal with big data, we will employ deep learning models which enables causal effect estimation using big data.
Candidates should have (or expect to achieve) a UK honours degree at 2.1 or above (or equivalent) in Computer Science or related disciplines with knowledge of Machine Learning; Python Programming; Software Engineering.
Formal applications can be completed online: https://www.abdn.ac.uk/pgap/login.php
• Apply for Degree of Doctor of Philosophy in Computing Science
• State name of the lead supervisor as the Name of Proposed Supervisor
• State ‘Self-funded’ as Intended Source of Funding
• State the exact project title on the application form
When applying please ensure all required documents are attached:
• All degree certificates and transcripts (Undergraduate AND Postgraduate MSc-officially translated into English where necessary)
• Detailed CV, Personal Statement and Intended source of funding
Informal inquiries can be made to Dr M Zhong ([Email Address Removed]), with a copy of your curriculum vitae and cover letter. All general enquiries should be directed to the Postgraduate Research School ([Email Address Removed])