CogniGron@Work Blog

Neuromorphic computing is rich in innovative components, but still lacks a clear blueprint for assembling them into large, functional systems. In this blog, Dr Madison Cotteret explains why deep learning methods fall short for spiking neural networks and why higher-level mathematical structure is needed. His work uses Vector-Symbolic Architectures (VSAs) to translate symbolic rules and relationships into robust, scalable neural populations. This approach works for both discrete systems, like finite state machines, and continuous structures such as spatial representations. The result is one of the first general frameworks for designing large-scale spiking architectures that remain hardware-friendly and noise-tolerant.

As silicon technology nears its physical limits, researchers at CogniGron are exploring bold new frontiers in computing. In our latest blog post, PhD candidate Ishitro Bhaduri introduces probabilistic computing—an approach that swaps stable binary bits for flickering p-bits, capable of tackling uncertainty and complex optimisation problems with ease. Using a remarkable material called lanthanum strontium manganite (LSMO), which naturally fluctuates between conducting and insulating states, his work offers a glimpse into a post-silicon future where randomness becomes a computing advantage. Discover how embracing probability could redefine the very way machines think.

What if one day, while walking across a bridge, you noticed a group of tiny robots crawling along its surface? They’re not pests — they’re protectors.In our second blog post, student assistant Amalie Fedorova talks with Bahar Haghighat, a CogniGron researcher exploring the fascinating world of swarm robotics — where small, intelligent robots work together to monitor structures like bridges and pipelines, keeping us safe in ways humans and static sensors simply can’t.Inspired by nature’s collective intelligence — think ants, birds, or fish — these robotic swarms move, sense, and think together. Equipped with vibration sensors and magnetic wheels, they can detect subtle changes that signal potential structural damage long before it becomes dangerous.They’re mobile, adaptable, and a glimpse into a future where machines quietly safeguard our infrastructure.

The first post of this blog series takes you inside the CogniGron Talent Network (CTN), a community launched in 2024 that brings together PhDs and postdocs from across different disciplines such as physics, AI, and materials science. Think monthly lectures that make neuromorphic concepts easy to grasp, lab tours, and yes, even bowling nights. Beyond the fun, CTN helps new PhDs find their footing, connects researchers across disciplines, and is now exploring career opportunities and industry collaborations.