Thanks BrainShit from the other forum. (Name is correct, LOL)
Papers like below are very important to validate Akida technology in various use cases.
In this case: Space
Ethics declarations:
This work was founded by the European Space Agency (contract number: 4000135881/21/NL/GLC/my)in the framework of the Ariadna research program. The authors declare that they have no known competingfinancial interests or personal relationships that are relevant to the content of this article.
16 May 2025
Energy efficiency analysis of Spiking Neural Networks for spaceapplications
Paolo Lunghi1(B), Stefano Silvestrini1, Dominik Dold2, **riele Meoni2,3, Alexander Hadjiivanov2 andDario Izzo2
1. Politecnico di Milano, Department of Aerospace Science and Technology, Via La Masa 34, 20156,Milano, IT
2. European Space Agency, Advanced Concepts Team, Keplerlaan 1, 2201 AZ Noordwijk, NL
3. European Space Agency, Φ-lab, Via Galileo Galilei 1, 00044 Frascati, Italy
Some small snippets:
Particular emphasis is placed on models based on temporalcoding, where crucial information is encoded in the timing of neuron spikes. These models promise evengreater efficiency of resulting networks, as they maximize the sparsity properties inherent in SNN. Areliable metric capable of comparing different architectures in a hardware-agnostic way is developed toestablish a clear theoretical dependence between architecture parameters and the energy consumption thatcan be expected onboard the spacecraft. The potential of this novel method and his flexibility to describespecific hardware platforms is demonstrated by its application to predicting the energy consumption of aBrainChip Akida AKD1000 neuromorphic processor.
A preliminary successfuldemonstration is given for the BrainChip Akida AKD1000 neuromorphic processor. Benchmark SNNmodels, both latency and rate based, exhibited a minimal loss in accuracy, compared with their equivalentANN, with significantly lower (from −50 % to −80 %) EMAC per inference. An even greater energyreduction can be expected with SNN implemented on actual neuromorphic devices, with respect tostandard ANN running on traditional hardware.
https://arxiv.org/pdf/2505.11418
(20min delay)