Evolutionary Computation and Machine Learning for Trustworthy and Inclusive Healthcare

Aim

Healthcare problems are noisy, high-dimensional, multi-objective and safety-critical, and the data behind them are spread across hospitals, devices and patients who cannot share them freely.

Machine learning (ML), and deep learning in particular, has driven much of the recent progress in medical imaging, biosignal analysis and clinical prediction. Evolutionary computation (EC) brings complementary strengths: it needs no gradients, handles conflicting objectives such as accuracy, fairness and interpretability, searches over discrete structures (feature subsets, programs, neural architectures, schedules) and, through genetic programming, can produce models that clinicians can read. Combining the two, through evolutionary machine learning, evolutionary AutoML and neural architecture search, surrogate-assisted optimisation and learning-guided search, suits health problems where data are scarce, heterogeneous and sensitive.

Health AI also has to work for everyone. Older adults, people with disabilities, children and other groups are under-represented in health data and are often poorly served by one-size-fits-all models, sensors and interfaces: an activity-recognition model trained on young adults may misread the gait of an older person, and a diagnostic app may be unusable for someone with low vision or limited dexterity. Algorithmic fairness, which asks that models perform equally well across groups, is one part of the answer. Inclusive health AI goes further and designs data collection, models and interaction around the needs of specific populations.

At recent CECs, healthcare papers have centred on evolutionary feature selection for diagnosis, neural architecture search for medical imaging, hospital scheduling and epidemic modelling. Some of the questions that matter most for deployment are still rare. How should we learn from decentralised, privacy-sensitive data, for example through federated learning across hospitals and wearables? How do we keep models fair across patient groups, serve specific populations such as older adults and people with disabilities, and stay robust when patients, devices and sites differ? How can EC and ML work with large language and foundation models in clinical decision support?

This special session brings together researchers in evolutionary computation, machine learning, federated and trustworthy AI, pervasive and accessible computing, along with clinical practitioners and researchers in ageing and disability. We welcome new EC methods, new ML methods and especially work that combines the two, as well as applications on real clinical or health data and new benchmarks. The session complements the long-running CEC series on Evolutionary Computation in Healthcare Industry, which focuses on operations problems, by focusing on data-driven, learning-centred, trustworthy and inclusive health AI.

Scope

We welcome papers applying evolutionary computation, machine learning, or both to health and medicine. Methods include evolutionary algorithms, genetic programming, swarm intelligence, multi-objective and surrogate-assisted optimisation, neuroevolution, deep learning, federated and transfer learning, reinforcement learning, fuzzy systems, and large language and foundation models. Topics include, but are not limited to:

• Evolutionary machine learning for healthEvolutionary AutoML, neural architecture and hyperparameter search, surrogate-assisted and learning-guided optimisation for medical models
• Federated and privacy-preserving health learningFederated learning and EC for federated learning across hospitals and devices: client selection, personalisation, multi-objective and federated NAS, decentralised and swarm learning
• Fairness and equity in medical AIMulti-objective trade-offs between accuracy and fairness across patient groups, sites and populations; bias in health data
• Inclusive health for specific populationsOlder adults (ageing in place, fall and frailty detection, cognitive decline), children (e.g., ADHD and autism assessment) and under-served or low-resource communities
• Accessibility and assistive health technologyAssistive and accessible systems for people with disabilities, adaptive interfaces, sign-language and alternative interaction for care
• Robustness and generalisationTransfer and domain adaptation across patients, devices and institutions; distribution shift; incremental learning
• Pervasive and wearable healthActivity recognition, sleep monitoring, cognitive and neurodegenerative assessment (e.g., Parkinson’s and Alzheimer’s disease)
• LLMs and foundation models for healthMedical foundation models, evolutionary prompt and program search, LLM-assisted clinical decision support, agent-based simulation of patients and care pathways
• Interpretable diagnosis and prognosisEvolutionary feature selection, GP-based feature learning and explainable ML for clinical, imaging, omics and text data
• Medical imaging and biosignalsDeep learning and evolutionary NAS for segmentation, classification and registration; EEG, ECG and EMG analysis
• Treatment and personalised medicineReinforcement learning and evolutionary optimisation for dosing, therapy and radiotherapy planning; dynamic treatment regimes
• Drug and molecule designMulti-objective molecular optimisation, peptide and protein design
• Public health and epidemicsEpidemic modelling, intervention and vaccine allocation
• Healthcare operationsScheduling, rostering, ambulance dispatching and hospital resource planning under uncertainty
• Serious games and interactive healthGames, gamification and adaptive interfaces for rehabilitation, therapy and health behaviour
• Benchmarks and real-world deploymentDatasets, benchmark problems, competitions and lessons learned from clinical deployment

Highlighted topics are the focus of this session and are rare in past CEC proceedings.

Important Dates

Dates follow the CEC 2027 website and may change. Please check there for the latest information.

Submission

Papers follow the standard CEC 2027 format and review process. Submit through the CEC 2027 submission system and select this special session as the topic. Author instructions and templates are on the CEC 2027 website. Special session papers follow the same format, page limits and review standards as regular papers, and accepted papers are presented at the congress and included in the main CEC 2027 proceedings.

The session will feature 15-minute paper presentations followed by a short panel on bringing evolutionary and machine learning methods into clinical practice.

Organisers

Related Reading

A selection of related work to give prospective authors a sense of scope. It is not exhaustive.

1. S. Tang, Q. Chen, B. Xue, M. Huang, M. Zhang. Genetic Programming with Multi-Task Feature Selection for Alzheimer’s Disease Diagnosis. CEC 2024. doi
2. Q. U. Ain, H. Al-Sahaf, B. Xue, M. Zhang. Exploring Genetic Programming Models in Computer-Aided Diagnosis of Skin Cancer Images. CEC 2024. doi
3. T. Dang, T. T. Nguyen, C. F. Moreno-Garcia, E. Elyan, J. McCall. Weighted Ensemble of Deep Learning Models based on Comprehensive Learning PSO for Medical Image Segmentation. CEC 2021. doi
4. L. Qu, C. Wang, Y. Shi. Brain Storm Optimization based Swarm Learning for Diabetic Retinopathy Image Classification. CEC 2024. doi
5. B. M. P. Silva, H. S. Bernardino, H. J. C. Barbosa. Human Activity Recognition Using Parallel Cartesian Genetic Programming. CEC 2021. doi
6. A. Sutradhar, A. A. Bidgoli, S. Rahnamayan. COR-MFS: A Correlation-Based Multi-Objective Feature Selection on EEG Signals. CEC 2024. doi
7. T. Hömberg, S. Mostaghim, S. Hiwa, T. Hiroyasu. Optimized Drug Design using Multi-Objective Evolutionary Algorithms with SELFIES. CEC 2024. doi
8. J. Sargant, S. Houghten, M. Dube. Evolving Weighted Contact Networks for Epidemic Modeling. CEC 2022. doi
9. J. MacLachlan, Y. Mei, F. Zhang, M. Zhang. Genetic Programming for Vehicle Subset Selection in Ambulance Dispatching. CEC 2022. doi
10. T. Bartz-Beielstein et al. Optimization and Adaptation of a Resource Planning Tool for Hospitals Under Special Consideration of the COVID-19 Pandemic. CEC 2021. doi

11. H. Zhu, Y. Jin. Multi-Objective Evolutionary Federated Learning. IEEE TNNLS, 2020. doi
12. H. Zhu, Y. Jin. Real-Time Federated Evolutionary Neural Architecture Search. IEEE TEVC, 2022. doi
13. Y.-J. Zheng, X. Chen, Q. Song, J. Yang, L. Wang. Evolutionary Optimization of COVID-19 Vaccine Distribution With Evolutionary Demands. IEEE TEVC, 2023. doi
14. J. Wei et al. Genetic U-Net: Automatically Designed Deep Networks for Retinal Vessel Segmentation Using a Genetic Algorithm. IEEE TMI, 2022. doi
15. Q. Zhang, J. Liu, Z. Zhang, J. Wen, B. Mao, X. Yao. Mitigating Unfairness via Evolutionary Multiobjective Ensemble Learning. IEEE TEVC, 2023. doi
16. Q. Zhang, J. Liu, X. Yao. Fairness-Aware Multiobjective Evolutionary Learning. IEEE TEVC, 2025. doi
17. Y. Chen, X. Qin, J. Wang, C. Yu, W. Gao. FedHealth: A Federated Transfer Learning Framework for Wearable Healthcare. IEEE Intelligent Systems, 2020. doi
18. Y. Chen, W. Lu, X. Qin, J. Wang, X. Xie. MetaFed: Federated Learning Among Federations With Cyclic Knowledge Distillation for Personalized Healthcare. IEEE TNNLS, 2023. doi
19. Z. Liu, Y. Chen, Y. Zhao, H. Yu, et al. Contribution-Aware Federated Learning for Smart Healthcare. AAAI 2022 (IAAI). doi
20. A. Z. Tan, H. Yu, L. Cui, Q. Yang. Towards Personalized Federated Learning. IEEE TNNLS, 2023. doi