- Spiking Neural Networks for next generation AI and its application
- 3D image processing, recognition algorithm development and its applications in intelligent HCI
- Neuromorphic-computing models (software & hardware) and its applications
- Artificial intelligence technology development for big data (e.g. finance, healthcare, education, etc.)
- Deep learning and dynamic modelling for continuous emotion recognition
- Deep learning for sematic image segmentation and its applications
- Human-like computing architecture and deep neural network modelling for real applications
- Real-time time series event detection and pattern recognition algorithm and its hardware implementation
- Internet of Things (IoT) system design and implementation
- Intelligent detection and monitoring systems for high speed railway applications
If you are interested in doing a PhD or Postdoc under my supervision related to any of the above topics, please feel free to send me an email with your CV and your brief proposal. Please also regularly check the PhD funding opportunities and other scholarships, studentships and bursaries at Brunel (It will be regularly updated). There is a special PhD scholarship (2+2) for Chinese students.
PhD projects for research students
In this project, advanced deep learning algorithms will be developed for medical data (e.g. biomedical signals, images, sound, touch, pressure, etc.). The sensors will be studied, signal processing methods will be applied for feature extractions, the dynamic of the data will be modelled and associated deep neural network models will be developed for clinical applications (e.g. detection, diagnosis, measurement, evaluation, etc.). The ideal candidate should have good knowledge on mathematics, signal processing, machine learning and sensors. Good programming skills are desirable.
The first generation of Artificial Intelligence (AI) research was rules-based and emulated classical logic to draw reasoned conclusions within a specific, narrowly defined problem domain. It was well suited to monitoring processes and improving efficiency. The second, current generation is largely concerned with sensing and perception, such as using deep-learning networks to analyse the contents of a video frame. A coming next generation will extend AI into areas that correspond to human cognition, such as interpretation and autonomous adaptation. Next-generation AI must be able to address novel situations and abstraction to automate ordinary human activities.
In this project, advanced next generation AI algorithms will be developed based on spiking neural network models for advanced human computer interaction systems. Spiking neuron models are from human neuroscientist after studying the architecture of the biological brain. These brain-like algorithms will simulate human brain functions like information encoding, storing, recall, reasoning and interfering. Complex connections and non-linear functions in the large scale neural network models can solve advanced AI problems. These algorithms will be used for human emotion recognition from multiple modalities such as facial expression, voice, music, touch, gesture, and biomedical signals like EEG signals. It will be then applied in advanced human computer interaction systems (e.g. robots, computer games, etc.) and health care applications. For example, the robots, equipped with emotion sensing, can understand human's behaviour and emotion, and can interact to human with human-like behaviour and emotion.
Emotion is a complex thing for human beings. It can reflect the mental and cognitive state of the person. Automatic detection of the emotions from facial expression, speech, written text in social media, body movement, gesture, bio signals such as brainwave (EEG), and heart-rate can provide accurate information about the mental state of the person. It can be used by doctors for clinical applications like depression, autism, pain, etc. It also can be used in intelligent human-computer interaction systems for robots, games and interactive films, etc.
In the project, we will develop deep learning-based artificial intelligent systems for recognising human emotions based on multiple modalities. The developed algorithms will be applied for intelligent human-computer interaction systems. It will be evaluated based on both accuracy and implementation speed for practical applications. The students with good mathematical knowledge (e.g. statistics, machine learning theory) and strong programming skills (e.g. MATLAB, Python, etc.) will be desirable.
Neuromorphic-computing models are the ones that can simulate human brains architecture and connections. It is composed of large numbers of neurons and nonlinear connections. The developed deep neural network models can do reasoning, pattern recognition, making decisions automatically and information retrievals like human brains. It will be base for the next generation of artificial intelligence. The students with good knowledge on mathematics, computer science, artificial intelligence, data analysis and advanced programming skills in Python, Java, MATLAB will be desirable.