MINDS researchers present at NEXT-AI 2026

Two MINDS doctoral researchers travelled to Loughborough in April 2026 to present their work at NEXT-AI; a national workshop bringing together researchers, industry representatives and exhibitors working on neuromorphic technologies and hardware-enabled AI.

Aiden Graham presented a poster examining how variation in the physical properties of memristor devices affects the classification accuracy of neural networks. Memristors are electronic components that can store and process information in ways that mimic the behaviour of biological synapses, making them a promising building block for AI hardware that is far more energy-efficient than conventional processors. Aiden's work asks a practical question that matters for anyone hoping to build this kind of hardware at scale: how much does device-to-device variation an unavoidable feature of physical fabrication affect the intelligence running on top of it?

Michail Rontionov presented NIR2FPGA, a compilation tool that takes a neuromorphic algorithm defined in a software simulator and automatically produces an energy-efficient hardware implementation on an FPGA. The tool bridges a significant gap between neuromorphic algorithm development and physical deployment, making it easier for researchers and engineers to move from a working software model to a working hardware system.

Both presentations reflect the MINDS CDT's core focus: developing AI that works not just in theory, but in the physical devices the world runs on.