Research programmes

In our Lab, we take a probabilistic approach to discovery AI, using Contextual Probability to enable discovery AI. Thus our work is categorised into two programmes: Contextual Probability and Discovery AI.

• Contextual Probability
  • ▸ What is it? Contextual Probability (CP) is a perceptionist probability. CP is an interpretation of probability that is distinct from the frequentist and Bayesian approaches. In this perception-based framework, the probability distribution P is transformed into a new probability G based on the context in which the probability is observed or perceived. Unlike frequentist probability, which relies on long-run frequencies, and Bayesian probability, which incorporates prior beliefs and evidence, contextual probability approximates P via contexts or neighbourhoods.
  • ▸ If you are interested, you can find out more at This Talk
  • ▸ (Exciting discoveries have been made recently. Stay tuned!)
• Discovery AI
  • ▸ What is it? Discovery AI refers to the area of artificial intelligence that is focused on uncovering new knowledge including novel insights, signatures (such as biomarkers), causal relationships, and hypotheses across various fields including healthcare, science, and law. Key components of Discovery AI include machine learning and deep learning, natural language processing, and causal inference.
  • ▸ Scope:
    • ○ Detection
    • ○ Causality
    • ○ Hypothesis - World Modelling
  • ▸ Application: AI for Science