Centre for Autonomous and Intelligent Systems (CAIS)

Knowledge-based AI

Knowledge-based AI develops methods for representing knowledge explicitly and using it to reason, explain conclusions and support decisions. Our expertise includes knowledge representation, ontologies and knowledge graphs; automated reasoning and computational argumentation; formal methods for specification and verification; and symbolic AI methods, including logic-based and algebraic approaches. These approaches allow concepts, relationships, assumptions, rules and constraints to be expressed in forms that people and machines can inspect and use.

Knowledge-based and data-driven AI are complementary rather than separate. Machine learning can help construct, complete and refine knowledge graphs, while explicit knowledge can guide learning and help ground and interpret its results. Natural language processing and argument mining can extract claims, evidence and relationships from text for subsequent reasoning. Fuzzy systems can combine expert-defined rules with data-driven learning, while robotics integrates knowledge and reasoning with learned perception and control. Hybrid and neuro-symbolic approaches bring these elements together, including using knowledge graphs, rules and formal verification to ground or evaluate generative AI. 

Data-driven AI

Data-driven AI develops models that learn from observations to identify patterns, estimate uncertainty and generate predictions, representations or content. Our expertise includes statistics, statistical modelling and inference, machine learning, deep learning and generative AI. These approaches support learning from structured and unstructured data and underpin research in prediction, classification, pattern discovery, language technologies and perception.

Many areas span both data-driven and knowledge-based AI. Natural language processing combines learned models with linguistic and semantic structure, while argument mining connects text analysis with computational argumentation. Knowledge graph learning applies statistical and neural methods to construct and complete structured knowledge. Fuzzy systems combine learning with graded concepts and rule-based reasoning. Robotics and autonomous systems integrate learned perception and control with explicit world models, constraints and reasoning. Hybrid and neuro-symbolic AI brings these elements together, including grounding generative AI in curated knowledge and using reasoning or verification methods to assess the consistency of its outputs.

Professor Mauro Vallati

School of Computing and Engineering
University of Huddersfield
Queensgate, Huddersfield
HD1 3DH

Email: m.vallati@hud.ac.uk

Healthcare

Artificial intelligence can support better health outcomes by helping professionals interpret complex information, identify patterns and make timely, well-informed decisions. CAIS brings together research that addresses healthcare challenges while recognising the importance of safety, transparency, fairness and human oversight. 

Our work encompasses disease detection, diagnosis and outcome prediction; clinical decision support; medical image analysis; remote monitoring and telehealth; bioinformatics and genomics; and the planning and management of healthcare services. Across these areas, we investigate how intelligent systems can complement clinical expertise, improve the use of data and support more personalised, accessible and efficient care. 

Healthcare applications must earn the confidence of patients, practitioners and wider society. We therefore consider not only technical performance, but also how systems can be interpreted, evaluated and deployed responsibly. By working across disciplinary boundaries, the Healthcare theme seeks to translate advances in AI into practical approaches that improve decision-making and quality of life. 

Leader: Tianhua Chen

Resilience

Resilience is the ability of organisations, communities, infrastructure and interconnected systems to prepare for disruption, withstand its effects, adapt and recover. CAIS investigates how artificial intelligence can help make complex systems more robust in the face of uncertainty, whether disruption arises from operational failures, natural hazards, economic shocks, climate-related pressures or wider societal change.

Our research addresses applications including supply chains, manufacturing, robotic systems, critical infrastructure, complex networks, geographic information systems and emergency response. We explore how AI can reveal dependencies and vulnerabilities, anticipate emerging risks, evaluate alternative courses of action and support timely, evidence-informed decisions before, during and after disruption.

Because resilience decisions can have significant consequences for people, organisations and public services, trustworthy use of AI is central to this theme. We place particular emphasis on transparent, interpretable and explainable systems that enable decision-makers to understand the evidence behind recommendations, exercise appropriate oversight and act with confidence.

Leader: George Bargiannis

Sustainability

The CAIS Sustainability theme examines both the sustainability of computing itself and the use of intelligent computing to support wider sustainability goals. Our research investigates how digital systems can be designed, built and operated with lower resource use and carbon impact, while remaining maintainable and evolvable without accumulating prohibitive technical debt. 

Our expertise includes energy-efficient and green computing; high-performance analytics for optimising data-intensive workflows on current and emerging platforms; sustainable software engineering; cloud computing; and data management. We also develop digital twins, using real-time data, simulation, machine learning and automated reasoning to create dynamic virtual representations of physical assets and systems, investigate alternative scenarios and support better-informed decisions. 

This work takes a whole-system view, addressing four interconnected dimensions of sustainability: the environment, the economy, society and individuals. By evaluating the total costs, impacts and benefits of computing technologies throughout their lifetimes, we aim to develop intelligent systems that contribute to sustainability without displacing technical, environmental or societal costs elsewhere. 

Leader: Richard Hill

Facilities and Resources

CAIS researchers use specialist facilities and shared digital resources to develop, test and evaluate artificial intelligence in realistic settings. These resources support research across autonomous systems, intelligent environments and complex supply networks, while creating opportunities for collaboration with academic and industry partners.

Robotics Laboratory 

Our Robotics Laboratory provides a practical environment for developing and evaluating autonomous and intelligent systems. Its facilities include Harold, a humanoid robot, and a range of robotic arms supporting research into areas such as perception, planning, reasoning, control, robotic manipulation and human-robot interaction. The laboratory enables researchers to investigate how intelligent systems can perceive their surroundings, make decisions and interact safely and effectively with people and physical environments.

Smart House 

The University’s Smart House is a two-storey, well-equipped research facility that provides a realistic environment for developing and evaluating smart technologies and services for the built environment. The facility supports research in areas including smart health, visual and audio analytics, energy management, intelligent environments and smart manufacturing. It enables AI technologies to be studied under realistic but controlled conditions and supports collaboration between researchers, businesses and public-sector organisations.

Supply Chain Data Hub 

CAIS is a sponsor of the Supply Chain Data Hub, an open platform created by the Supply Chain AI Lab at the University of Cambridge. The Hub brings together open datasets, curated research, AI tools and international data challenges at the intersection of AI and supply chain management. It provides researchers and practitioners with resources for developing and evaluating new approaches to understanding complex supply networks and building more resilient and adaptive supply chains. Through its sponsorship, CAIS supports open research infrastructure and collaboration across the international Supply Chain AI community.