The Neuromorphic Edge Computing Systems Lab strives to create computing algorithms, circuits, systems, and architectures for emerging applications and a more sustainable computing scenario.

Our long-term research goal is to enable intelligence inside our microchips by designing neuromorphic circuits and systems that mimic the brain's fundamental information processing strategies for solving challenging real-world problems.

We aim to build brain-inspired machine intelligence devices. We address the problem of machine intelligence across the whole computing stack, from new models of computation down to hardware. We take inspiration from the brain's efficiency, and we research neural-inspired models of computation that are massively parallel, compute on-demand, and benefit from emerging nano- and microelectronics technologies to develop new disruptive neuromorphic computing systems.

The specific research activities in the NECS Lab focus on embedding intelligence in edge devices in the form of adaptability, robustness, long-life and online learning, intent, and explainability. We research performance optimization and energy tradeoffs for specific machine intelligence tasks by co-design algorithms and hardware. We develop complementary metal-oxide-semiconductor (CMOS) Very Large Scale Integrated (VLSI) neuromorphic devices with in-memory and near-memory computing abilities to replicate bio-realistic spiking neural networks interconnected with learning synapses.

Applications 

Emerging applications in brain-computer interfaces, health monitoring, intelligent robotics, and the internet-of-things require increasing computing power in tiny chips that need to work with as little energy as possible and need to process increasingly more data. Using traditional performance scaling methods, scaling at the lowest levels of the hardware as the microelectronic industry has been doing in the last decades, we can still improve, but not enough. In the long term, to go beyond the limits of physical scaling, we need to research disruptive technology solutions together with new models of computation.