A Swansea University researcher whose work could help create more inclusive icons and symbols for neurodivergent people has been awarded a prestigious scholarship to present her findings at a leading international psychology conference.

Researchers at Zhengzhou University have developed a Janus bioelectrode that protects weak body signals from electromagnetic interference. By combining a magnetic-dielectric aerogel with a liquid-metal conductive textile, the electrode records high-quality muscle and heart signals even in noisy electronic environments. When paired with machine learning, it enabled accurate recognition of human motions using fused EMG and ECG signals.

- Professor Min-Soo Kim’s research team in the School of Computing develops RL-SPH, a reinforcement-learning-based method that independently generates feasible solutions without relying on an external optimization solver to enforce feasibility

- Achieves 100% feasibility across five representative ILP benchmarks, with an average 28.6× lower primal gap than fully feasible baselines

- Trains 14.7× faster than representative learning-based approaches, demonstrating potential for AI-based decision-making in logistics, manufacturing, semiconductor production, and workforce management