Skip to main content
Yuhua Li

Dr Yuhua Li

Reader of Machine Learning

School of Computer Science and Informatics

+44 29208 75317
Abacws, Room 4.59, Senghennydd Road, Cathays, Cardiff, CF24 4AG
Available for postgraduate supervision


Data Analytics and Machine Learning Research Group

I have conducted fundamental and applied research in machine learning, pattern recognition, data science, semantic similarity analysis and condition monitoring.

My experience in machine learning and pattern recognition includes statistical, geometrical methods and neural networks for feature/pattern selection and data analysis, knowledge discovery and inference.

My contribution to machine learning includes the development of anomaly/novelty detection methods for safety/mission-critical systems, which have limited or no data/knowledge on rare events, and informative observation selection techniques for sensors/measurements location optimisation for problems such as effective monitoring and process control. My works have motivated other researchers to develop new AI algorithms, use them as benchmarks and adopt them in products.

I have led and carried out research projects funded by the government and industry. I have collaborated on research projects with different sizes of national and international companies. My research has been applied to solve problems in digital manufacturing, condition monitoring, financial engineering, and other real-world problems.

External Engagement

  • Member of the EPSRC Peer Review College
  • Academic Adviser to the Commonwealth Scholarship Commission UK
  • Member of the Innovation Advisory Council for Wales
  • Associate Editor of the IEEE Transactions on Neural Networks and Learning Systems



























Data Analytics and Machine Learning Research Group

My research interests include:

  • Machine learning, pattern recognition
  • Novelty detection, anomaly detection
  • Data science, Big Data, text mining
  • Neural networks, deep learning
  • Hyperdimensional computing, vector symbolic architectures
  • Condition monitoring and signal processing
  • Machine learning and AI applications, e.g., cyber secuirity, finance, manufacturing

Selected publications (more publications on Google Scholar).


I received a postgraduate certificate in higher education, I am a Fellow of the HEA. To teach:

  • CMT307 Applied Machine Learning
  • CMT316 Applications of Machine Learning: Natural Language Processing/Computer Vision
  • CMT219 Algorithms, Data Structures and Programming
  • CM1210 Object Oriented Java Programming


Current PhD students

I am the 1st supervisor for: 

I am interested in supervising PhD students in the areas of:

  • Machine learning, pattern recognition
  • Data science, Big Data, text mining
  • Neural networks, deep learning
  • Hyperdimensional computing, vector symbolic architectures
  • Machine learning and AI applications, e.g., cyber secuirity, finance and engineering

You are welcome to contact me ( if you have an outstanding academic background and high ambition for research excellence. Listed below are examples of PhD project proposals.

Project 1: Machine Learning for Health Monitoring of New-born Babies in Hospital Intensive Care Units

Hospital Neonatal Intensive Care Units (NICU) employ medical devices and sensors to continuously collect physiological data such as heart rate, temperature, and ECG. This project aims to harness the wealth of data available in NICU to develop effective solutions for neonate monitoring and early detection of diseases, with a particular focus on sepsis, which is responsible for 26% of neonate deaths. Utilizing machine learning techniques, the project will extract features from data streams, identify risk factors, and detect disease early to assist clinicians in providing better care for NICU patients. The project will be carried out in partnership with the School of Medicine, using NICU data collected from hospital wards.

Project 2: Transforming Machine Intelligence for Resource-Constrained Applications

In recent years, machine learning systems, including deep learning, have achieved remarkable success in various applications. However, they face significant challenges, such as high energy consumption, adapting to new conditions, and data deficiency, such as missing values, data imbalance, and data privacy. To address these challenges, this research harnesses the latest advancements in Vector Symbolic Architectures (VSA) or Hyperdimensional Computing (HDC), an emerging computing framework, to develop light-weight algorithms that can efficiently perform cognitive processing for resource-constrained applications at the network edge. VSA is inspired by the brain and represents and manipulates data in a high-dimensional vector space. Its holographic distributed representation and manipulation of information make computing more robust to noise, scalable, energy-efficient, and require less time and data for training and inference. This project aims to develop resource-efficient and data-frugal algorithms that can enhance and complement existing machine learning and deep learning methods.

Project 3: Online classification with emerging new classes

Standard classification methods can only classify pre-defined classes, i.e., they classify a new instance into one (or multiple) of the known classes. For example, for building a classifier for viral respiratory diseases, we need to train the classification model on a dataset with pre-defined classes such as MERS and SARS. At the time of developing a model for disease classification, the classifier is trained on available data which contains only, e.g., MERS and SARS. Such a classifier can only classify MERS and SARS diseases, it will be unable to deal with the emergence of new diseases such as COVID-19 in the future. To deal with the emergence of new classes, a novel approach is needed to learn a classifier that is able to detect newly emerging classes and adapt the classifier accordingly. Such a classifier learning paradigm with new classes has numerous applications, e.g., self-driving cars manoeuvring in novel traffic scenarios, malware detectors dealing with new types of network attacks, robotic soldiers navigating in new types of terrains, etc.

This project aims to develop a novel approach to learning a classifier that is capable of classifying emerging and novel classes. The proposed approach will address two main challenges: effective detection of emerging classes and just-in-time adaptation of classifiers for new classes. Emerging class detection will be built on the latest advances in novelty detection (novelty detection is a machine learning technique that learns a model based on only known classes to detect instances coming from a novel class), and just-in-time adaptation will develop a novel incremental learning strategy to integrate new classes into the current classifier. The developed algorithms will be evaluated on a use case in cybersecurity or the Internet of Things (e.g., new types of network attacks).

Project 4: Learning concept evolution in data streams
In applications with concept evolution, new concepts emerge in the data stream, and existing/known concepts disappear over time, e.g., new types of attacks in a computer network and new topics of interest in a social media data stream. This project aims to develop novel methods for tackling the challenging issue of concept evolution to enable the learned models to accommodate new concepts. It will achieve the following objectives: known concepts modelling; novel instances detection and accumulation; new concepts detection and integration; outdated concepts retiring.

Project 5: Explainable machine learning for securing Internet of Things (IoT) 
Internet of Things (IoT) consists of things, services, and networks, it connects interrelated smart devices, objects, animals or people to transfer data over a network to serve people better. The amount of data transferred with IoT systems is continuous, heterogenous and huge, which makes IoT systems more vulnerable than the traditional network to malicious activities from attackers, so the security and privacy of this highly automated network is a key challenge for the deployment of Internet of Things (IoT). It is constantly subject to adversarial attacks including denial of service, jamming, spoofing, eavesdropping, malware and privacy leakage. The limited resources (computation, battery, and memory) on IoT devices and the amount of data generated and communicated severely constrain the applicability of existing security measures to IoT systems. Even if a security system is effective at the time of deployment, it is prone to fail soon as attackers adapt smarter strategies to foil the system and avoid detection. Machine learning is a major tool for detecting adversarial attacks, and human-level explainability of detection results remains open to research in the security of IoT. 

This project aims to address these key challenges to secure future IoT systems with creative machine learning methods by Investigating data streaming classification methods for effectively detecting known types of attacks and their variants in the future; Developing computationally cheaper machine learning algorithms as well as robustness against eavesdropping attacks; Optimising the offloading policy in dynamic radio environments to optimally distribute the computational load over cloud, device and edge; Investigating adversarial machine learning techniques to tackle attackers' changing strategies; Interpreting prediction results to support human to take trustworthy actions.