VIP ChallENG research goals
We will develop various social robot applications with multi-modal AI technologies such as knowledge representation and reasoning, machine/deep learning, computer vision and natural language processing. To do this we will need to investigate the analytical, cognitive and social responsibility aspects to create an integrated, functioning system.
You can choose one of these three aspects as the objective of your project:
Socially analytical AI
- You will develop deep learning methodologies for image- and video-based human behaviour analysis, such as trajectory prediction, social relationship recognition and behaviour prediction that are essential to facilitate intelligent human-robot interaction.
- For example, trajectory prediction is about predicting people’s walking trajectories in a natural scene, which need to follow socially and physically acceptable conventions. This is a very challenging problem and an essential component in many autonomous platforms such as self-driving cars and social robots. We will investigate the existing deep-learning approaches and develop advanced methods for this emerging problem.
Socially cognitive AI
- You will develop multi-modal AI methodologies to support social robot functionalities, e.g., keeping a healthy routine for elderly in nursing homes and reminding people to maintain social distancing in the workplace or shopping centres during pandemics like COVID-19.
- For example, to regulate social distancing, an integrated system can be developed comprising a computer vision-based monitoring component for crowdedness detection, a logical reasoning component to determine the appropriate occasions to intervene and a chatbot for effective delivery of advice to people to maintain social distancing. We will focus on the software application development to enable a well-functioning social robot.
Socially responsible AI
- People must be able to trust and be willing to engage with social robots in meaningful ways. To achieve this, you will research ethical theories and machine ethics, and develop AI programs to ensure social robots adhere to ethical principles and behavioural norms.
- For example, social robots in the workplace need to follow social conventions such as respecting people’s privacy and diversity. To enable such moral behaviours, we will develop a knowledge representation and reasoning system to first encode the ethical principles in a knowledge base and then perform reasoning to derive appropriate behaviours in different situations. Natural language processing will also be investigated for knowledge base generation.
Research, design or technical ChallENG
- Data collection and annotation
- Investigation of state-of-the-art AI methodologies
- Exploration of new AI methodologies
- Comparative study of different methodologies
- Development and experimental evaluation
- Quantitative and qualitative result analysis