FindAPhD Weekly PhD Newsletter | JOIN NOW FindAPhD Weekly PhD Newsletter | JOIN NOW

Detecting Anomalies from Space Data Sources using Deep Learning Data Analytics

   Faculty of Computing, Engineering and the Built Environment

This project is no longer listed on and may not be available.

Click here to search for PhD studentship opportunities
  Dr Y Bi  Applications accepted all year round  Funded PhD Project (European/UK Students Only)

About the Project

Analysing and then pinpointing anomalies in data sequences recorded by satellites observing our Planet from outer space require to discover change points in the data sequences. These change points can be formulated to be normal and abnormal changes, which are in turn referred to as anomalies. In our previous studies, the four methods: wavelet, statistical martingales, fuzzy-inspired and weighted local outlier factors, have been explored and the respective algorithms have been developed. The results found underpin new strategies for detecting change points within data sequences and will lead to further development of specialised deep learning detection algorithms. 

This project aims to develop change detection analytics mainly for data sequences predicted from observed data by a recurrent neural network or transformer network. The proposed work will involve development of deep learning methods to generate prediction models via incorporating temporal information embedded in the observed data and conduct investigations into two distributions: one estimated from the observed data, which will be used to train a prediction model, whereas another generated by the prediction model, which will be used for change detection. The two distributions will be studied in a supervised manner and then will be generalized to unsupervised tasks. Both supervised and unsupervised tasks would be treated as adversarial processes, which could result in a pragmatic approach for identifying potential changes in future observations. 

The algorithms developed will be employed for various real-world application scenarios, for instance, to analyse electromagnetic data collected by satellites of the European Swarm satellite constellation and the China Seismo-Electromagnetic Satellite (CSES) and establish correlation between abnormal changes and hazardous events such as earthquake or power generation and consumption from renewable energy sources. 

In addition to regular supervision, the successful candidate will be provided with a range of training opportunities, including courses provided by the University Doctor College, locally hosted seminar series with national and international experts organized by the School, and collaboration workshops with our research partners. We also encourage the candidate to participate in relevant symposia and conferences and discuss research findings with the wider audience in the scientific community.

Essential Criteria

Upper Second Class Honours (2:1) Degree in Computing Science, Mathematics, Electronic Engineering, Remote Sensing, Earth Observation or a cognate area

MSc degree (70%, first class) in the related areas.

Funding and Eligibility 

The scholarships will cover tuition fees at the home (EU) rate and a maintenance allowance of £15,609 per annum for three years. 

Applicants should have a nationality of an European Space Agency (ESA) member state (i.e. Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, The Netherlands, Norway, Poland, Portugal, Romania, Spain, Sweden, Switzerland, the United Kingdom) or an European Cooperating State (i.e. Latvia, Slovenia Estonia and Hungary) to be eligible for both fees and maintenance. 

This PhD opportunity is immediately available and will be open until the post is filled. For anyone who would be interested in the opportunity and the details of how to apply, please contact Dr Yaxin Bi in the following email address.

Dr Yaxin Bi

Reader in Artificial Intelligence

School of Computing

Ulster University at Jordanstown


County Antrim, BT37 0QB

United Kingdom

T: +44 2890 366582

E: [Email Address Removed] 


Funding Notes

European Space Agency


X. Kong, Y. Bi, and D. H. Glass. Anomaly Detection in Sequential Data Based on Subsequence Identification. Journal of Artificial Intelligence Review, 53(1): 625-652, 2020.
V. Christodoulou, Y. Bi, G. Wilkie. A tool for SWARM satellite data analysis and anomaly detection. PloS ONE, 14(4):e0212098 , April 2019
V. Christodoulou, Y. Bi and G. Wilkie. Seismic Anomaly Detection Using Symbolic Representation Methods, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 11(9): 3366-3397, 2018. DOI: 10.1109/JSTARS.2018.2854865
X. Kong, Y. Bi, and D. H. Glass. Detecting Seismic Anomalies in Outgoing Long-Wave Radiation Data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 8 (2): 649 - 660, 2015. DOI: 10.1109/JSTARS.2014.2363473
Y. Bi. Sentiment classification in social media data by combining triplet belief functions. The Journal of the Association for Information Science and Technology (JASIST).
PhD saved successfully
View saved PhDs