EBRAINS researchers present latest neuromorphic computing advances at NICE 2026

Researchers from the EBRAINS community presented their work at the 2026 Neuro Inspired Computational Elements (NICE) Conference in Atlanta (USA), on March 24-26 (agenda). Their contributions showcased recent advances in neuromorphic hardware, with all three presentations highlighting different aspects of the BrainScaleS-2 platform. Three conference papers have now been published in the IEEE conference proceedings. 

The BrainScaleS-2 system is a neuromorphic platform with versatile applications in the fields of computational neuroscience and spike-based machine learning. Researchers have free access to the system via EBRAINS. 

Real-Time Processing of Analog Signals on Accelerated Neuromorphic Hardware 

Yannik Stradmann, Johannes Schemmel, Mihai A. Petrovici and Laura Kriener 

This presentation demonstrated real-time processing of analogue signals on the BrainScaleS-2 neuromorphic platform, introducing direct analogue sensor input to the hardware without conventional signal conversion. The work showed how the system can localise sound sources in real time and directly control a servo motor, enabling a fully on-chip processing pipeline from sensory input to physical action. 

The BrainScaleS-2 Multi-Chip System: Interconnecting Continuous-Time Neuromorphic Compute Substrates 

Joscha Ilmberger and Johannes Schemmel 

One of the contributions introduced the BrainScaleS-2 multi-chip system, describing a scalable architecture that interconnects multiple neuromorphic compute substrates through FPGA-based communication. The system achieves sub-microsecond chip-to-chip latencies while enabling larger-scale neuromorphic computing deployments. 

Amortized Inference of Neuron Parameters on Analog Neuromorphic Hardware 

Jakob Kaiser, Eric Müller and Johannes Schemmel 

At another talk, researchers presented a new approach to amortized inference of neuron parameters on analogue neuromorphic hardware. Using simulation-based inference, the researchers demonstrated an efficient method for estimating the parameters of analogue neuron circuits on BrainScaleS-2, providing a valuable tool for calibrating and programming neuromorphic hardware. 

The slides and the video of this talk, as well as materials from other sessions, are available on the NICE agenda webpage. 

Hands-on tutorial  

On the tutorial day of NICE, the BrainScaleS team gave a hands-on tutorial, in which participants were able to explore BrainScaleS-2, one of the world’s most advanced analogue platforms for neuromorphic computing.  For the tutorial, the (machine-learning targeting) PyTorch and the (neuroscience targeting) PyNN-based software interfaces were shown. This allowed participants to gain insights into the unique properties and challenges of analogue computing and to exploit the versatility of the system by exploring user-defined learning rules. Each participant had the opportunity to follow a prepared tutorial or branch-off and implement their own project on the systems. The tutorial notebooks and access to the systems are available via https://ebrains.eu/nmc.