CAIH Lab
Research
Our methodological research centres on causal and agentic AI for personalised decision-making, with healthcare as a principal application domain.
Core areas
Research themes
Causal AI
Developing AI methods that reason about causes, interventions and counterfactuals.
Agentic AI
Exploring autonomous and collaborative AI agents for research and healthcare.
Personalised Healthcare
Developing data-driven methods for patient-specific treatments and clinical decision support.
Technical breadth
Wider research scope
- Personalised decision-making and individualised treatment-effect estimation
- Causal inference and discovery from large-scale observational data
- Counterfactual reasoning for fairness, explainability and decision support
- Causal foundation models and digital twins
- Decision-making under uncertainty and conformal prediction
- Agentic and collaborative AI systems
- Multimodal, federated and continual learning
- Synthetic data, causal benchmarking and evaluation
- Robustness, domain adaptation and out-of-distribution detection
- Applications in personalised healthcare and clinical decision support
Publications
A more up-to-date list may be available on Google Scholar.
Citation counts are provided by OpenAlex and Semantic Scholar and may differ between services. Altmetric reports online attention rather than citations. Citation data last updated: 5 October 2026.
2026
A Practical Tutorial on Agentic AI
Position: Causal AI for Personalised Treatments Needs Holistic Approach
Precision psychiatry: thinking beyond simple prediction models, without dismissing them
2025
Clinical prediction of Atrial Fibrillation after Cardiac Surgery: the AFACS risk scores
A context-aware methodology for fault detection and severity classification under variable operating conditions
Sample Selection Bias in Machine Learning for Healthcare
Individualised Treatment Effects Estimation with Composite Treatments and Composite Outcomes
Beyond Correlations: The Necessity and the Challenges of Causal AI
2024
GTAGCN: Generalized Topology Adaptive Graph Convolutional Networks
HCR-Net: A transfer learning based script independent handwriting character recognition network
Trolley Optimisation for Loading Printed Circuit Board Components
Temporal Dynamics Unleashed: Elevating Variational Graph Attention
CliqueFluxNet: Unveiling EHR Insights with Stochastic Edge Fluxing and Maximal Clique Utilisation using Graph Neural Networks
Continuous Patient State Attention Model for Addressing Irregularity in Electronic Health Records
Comparing the risks of new-onset gastric cancer or gastric diseases in type 2 diabetes mellitus patients exposed to SGLT2I, DPP4I or GLP1A: a population-based cohort study
A Brief Review of Hypernetworks in Deep Learning
Dynamic Inter-treatment Information Sharing for Individualized Treatment Effects Estimation
2023
Indic Script Family and Its Offline Handwriting Recognition for Characters/ Digits and Words: A Comprehensive Survey
Predicting supply chain procurement bid prices with uncertainty quantification and machine learning: a case study in aerospace manufacturing
A network science approach to identify disruptive elements of an airline
Multi-tier material consolidation in complex supply chains
Real-time large-scale supplier order assignments across two-tiers of a supply chain with penalty and dual-sourcing
Synthesizing Mixed-type Electronic Health Records using Diffusion Models
Improving Diagnostics with Deep Forest Applied to Electronic Health Records
Using Bottleneck Adapters to Identify Cancer in Clinical Notes under Low-Resource Constraints
Adversarial De-confounding in Individualised Treatment Effects Estimation
1567P The effect of SGLT2i and DPP4i on new-onset gastric cancer and gastric diseases in type 2 diabetes mellitus: A population-based cohort study
2022
COPER: Continuous Patient State Perceiver
A data-driven method to assess the causes and impact of delay propagation in air transportation systems
2021
Stochastic Optimization for Large-scale Machine Learning
Online handwritten Gurmukhi word recognition using fine-tuned Deep Convolutional Neural Network on offline features
LIBS2ML: A Library for Scalable Second Order Machine Learning Algorithms
2020
A Self Controlled RDP Approach for Feature Extraction in Online Handwriting Recognition using Deep Learning
The relationship between nested patterns and the ripple effect in complex supply networks
Stochastic Trust Region Inexact Newton Method for Large-scale Machine Learning
2019
Solving large scale linear support vector classification using an optimization framework based on stochastic approximation and coordinate descent approaches
Problem Formulations and Solvers in Linear SVM: a Review
SAAGs: Biased Stochastic Variance Reduction Methods for Large-scale Learning
2018
Faster Learning by Reduction of Data Access Time
2017
Trust Region Levenberg-Marquardt Method for Linear SVM
Mini-batch Block-coordinate based Stochastic Average Adjusted Gradient Methods to Solve Big Data Problems
2016
Online Support Vector Machine Based on Minimum Euclidean Distance
No publications match the selected filters.