An arXiv paper describes a co-designed spiking neural network for online continual learning on Intel Loihi 2. The abstract frames the problem: edge AI must adapt to new classes and shifting data within a tight power budget, yet online learning—updating from each sample before the next arrives—is where conventional processors are least efficient.
The authors propose a spiking neural network co-designed with the Loihi 2 neuromorphic chip to address this inefficiency. The abstract does not include experimental results or specific performance figures, so the claims are limited to the proposed approach and its intended context.
This work is relevant to edge AI applications where power constraints make conventional hardware impractical for continual adaptation. The paper is available on arXiv under identifier 2511.01553.