Computing Science and Mathematics Research

Data science And Intelligent Systems (DAIS)

About | News | Members, Expertise and Applications | Projects and Publications

The Data science And Intelligent Systems (DAIS) Research Group are a group of academics, researchers and PhD students within the Division of Computing Science and Mathematics at the University of Stirling. We are interested in the development, theory, and application of data-driven methods to understand, model and optimise the world.

Data is all around us, and humanity generates over 2 quintillion bytes (that’s a 2 followed by 18 zeroes) of new data every day. Cutting-edge techniques are able to make sense of this data and explore it for the greater good. DAIS researchers have embraced data to: inform healthcare decisions through modern statistical analysis; assist those with impairments communicate, using machine learning algorithms to spot c at the momenomplex patterns in sound and brain patterns; predict outcomes of complex systems, with intelligent models that explain their decisions so that they can be trusted; perceive the world through computer vision to assist with sight loss and monitor our environment from space; schedule planes to arrive on time and design buildings to be more comfortable and energy efficient through advanced optimisation methods; and understand how all these techniques work and fit together using novel visualisations methods. These are just some of the areas where our group is advancing knowledge.

We are always happy to talk to companies and other academics about how we might work together. In the first instance please contact Dr Sandy Brownlee (alexander.brownlee@stir.ac.uk), DAIS group leader.

News

May 2024: Industrial showcase, and Grant Success

The DAIS group hosted an industrial engagement event on 26 April, organised by Dr Jason Adair and Dr Sandy Brownlee. The event included success stories of academic/industrial partnerships, funding opportunities, and a poster session with lots of time for discussions between academics and visitors around potential projects. There was a real buzz about the day: around 15 external organisations participated with several follow-up meetings planned. We look forward to seeing what new partnerships emerge!

Dr Sandy Brownlee and Dr Leonardo Bezerra are pleased to share that they have been successful in winning a £410k grant from Responsible AI UK, part of the £3.5M project "Participatory Harm Auditing Workbenches and Methodologies" (PHAWM) being led by the University of Glasgow. Stirling's part is largely focused on exploring fairness and bias in machine learning algorithms. A full news story can be found here.

March 2024: Another viva success; and announcing our Industrial Showcase

Congratulations to Dr Stefano Sarti who passed his PhD viva with minor corrections this month. His thesis is entitled "Neuroevolution Trajectory Networks: Illuminating the Evolution of Artificial Neural Networks".

We would also like to announce our Industrial Showcase. Are you a business interested in benefiting from our cutting edge research in data science? We are holding a showcase day highlighting successful collaborations and how our expertise in machine learning, optimisation, and more can support you. Join us at the University of Stirling on Fri, 26 Apr 2024 10:00 - 13:30 to explore potential collaborations, funding & networking. Sign up here.

The Special Issue on Explainable AI in Evolutionary Computation on ACM Transactions on Evolutionary Learning and Optimization is finally online! This Special Issue collects four papers on the intersection between XAI and Evolutionary Computation, a growing field that still deserves more effort from the research community. Dr Sandy Brownlee is one of the guest editors on this issue.

January 2024: Viva success and more welcomes

Congratulations to Dr Penny Johnston who passed her PhD viva with minor corrections this month. Her thesis is entitled "Designing a Neuro-Symbolic Architecture for the Visual Question Answer Task".

We also welcome Dr Paulius Stankaitis and Dr Paul McMenemy as lecturers joining the group.

December 2023: Presentation

Dr Sandy Brownlee presented a paper on using large language models to automatically improve code at SSBSE 2023. This also attracted some media coverage here and here.

November 2023: More welcomes, and media coverage

This month we welcome Dr Vahid Akbari as a new lecturer. Vahid specialises in machine learning and earth observation technology and has previously worked within our Division as a postdoctoral research assistant. In other news, Dr Sandy Brownlee was recently interviewed about AI in Scotland. You can read the article here.

October 2023: Welcomes

Our department is currently growing and we are fortunate to have been joined by several new staff in recent weeks. Dr Ahsan Adeel brings experience of biologically plausible neural nets, and Dr Leonardo Bezerra has expertise in heuristic optimisation. You can see their keywords on our group members page. Welcome to both! We also look forward to more new staff joining during the next few months.

September 2023: Conferences, industrial visit, and farewells

In July five DAIS members travelled to Lisbon to attend the Genetic and Evolutionary Computation Conference (GECCO), the top conference in this topic. Gabriela Ochoa, Sandy Brownlee, Saemundur Haraldsson, Sarah Thomson, and Stefano Sarti were kept very busy! In addition to all presenting their research, Gabriela, Sandy and Saemi were giving tutorials in fitness landscape analysis and genetic improvement of software; Gabriela, Sarah, and Sandy were chairing workshops in fitness landscapes and explainable AI; and Gabriela won a best paper prize.

In September, Sandy Brownlee and MPhil student Giancarlo Catalano travelled to BT's research offices at Adastral Park, Ipswich. Gian is working on developing explainable optimisation algorithms for BT's staff rostering problems, and was reporting on initial results, which were well received.

Farewell and thanks to Jefersson dos Santos and Sarah Thomson, who have moved to new positions at the University of Sheffield and Edinburgh Napier University respectively. They have both made strong contributions to the research in the group and department and will be missed! Good luck in your new roles.

June 2023 (part 2): Research visits

Multiple travels this month. Dr Sandy Brownlee spent some time at the School of Architecture, Building, and Civil Engineering at Loughborough University in his role as a Visiting Fellow there. Several existing strands of work in optimisation of energy efficient buildings were developed, as well as some possible new links in transportation modelling and optimisation.

Dr Sarah Thomson recently visited Professor Sébastien Verel at Université du Littoral Côte d'Opale (ULCO) in Calais, France. There, she had many productive and inspiring collaboration and brainstorming meetings focused on the topic of fitness landscape analysis in evolutionary computation (although fractals and convolutional neural networks were involved too!). The trip strengthened the existing collaborative relationship between Stirling and ULCO in the field of evolutionary computation, and will probably catalyse the production of associated publications. During Sarah's time in France, she was also invited to a dinner in Lille with fellow DAIS member Prof Gabriela Ochoa and other members of the international evolutionary computation community. Gabriela was contributing to the summer school on Automatic Algorithm Design, and was recently also invited to participate in a Dagsthul seminar on Challenges in Benchmarking Optimization Heuristics.

Meanwhile Dr Simona Hapca recently hosted several staff from NHS Forth Valley to learn about statistical approaches and explore opportunities for research collaboration. A busy month!

June 2023: Welcomes, and Best Paper Award Nomination

Welcome to two new research students starting under Dr Sandy Brownlee's supervision: Hiba Alansari (co-supervised by Dr Jason Adair) is undertaking a Professional Doctorate on modelling and optimisation of solar panel output; Giancarlo Catalano (co-supervised by Dr David Cairns) is taking an MPhil in explainability of non-deterministic solvers. Prof Gabriela Ochoa has also been nominated for another award, for her paper "Pareto Local Optimal Solutions Networks with Compression, Enhanced Visualization and Expressiveness" (coauthors: Arnaud Liefooghe, Gabriela Ochoa, Sébastien Verel and Bilel Derbel) at this year's GECCO conference.

April 2023: Best Paper Awards

PhD student Stefano Sarti and his supervisors Prof Gabriela Ochoa and Dr Jason Adair won two awards for their paper "Under the Hood of Transfer Learning for Deep Neuroevolution" at this year's EvoStar conference.

March 2023: Congratulations

Mila Goranova, being supervised by Prof Gabriela Ochoa, passed her PhD viva this month. Her thesis investigated applications of evolutionary algorithms to antibiotic treatment regimes. Gabriela is also listed by Research.com in their 2023 Best Computer Scientists Rankings. Well done both!

January 2023: New Project

Sarah Thomson and Sandy Brownlee have been awarded a short consultancy project from NHS Argyll & Bute to optimise their pool vehicle fleet and support their move towards electric vehicles longer term. They have also recently had the news that the fitness landscapes and explainable AI workshops that they respectively lead at GECCO have been accepted for the 2023 conference.

November 2022: Keynote

Professor Gabriela Ochoa participated as a keynote speaker at the 8th IEEE Latin American Conference on Computational Intelligence (IEEE LA-CCI 2022), in Montevideo, Uruguay, 22-25 November, with a talk entitled "Neuroevolution Landscapes and Trajectories". Neuroevolution combines two key computational intelligence techniques: evolutionary computation and artificial neural networks. The idea is to use evolutionary algorithms to automatically design and optimise neural networks when solving complex computer vision, language processing or control problems. The talk covered her recent work on network-based visualisation and analysis techniques applied to this exciting domain.

October 2022: Publications and Conferences

Gabriela Ochoa and Sarah Thomson recently presented a paper at the Parallel Problem Solving in Nature Conference (PPSN 2022) in Dortmund. Dr Saemi Haraldsson recently has three papers accepted, two in the top-tier journal Transactions on Software Engineering, and one at the highly respected ESEC-FSE conference, all arising from his involvement in the Exploiting Defect Prediction for Automatic Software Repair (Fixie) project. Dr Sandy Brownlee has also had a paper accepted to IEEE Transactions on Fuzzy Systems on map matching for aircraft movements as part of his work in the project TRANSIT: Towards a Robust Airport Decision Support System for Intelligent Taxiing.