Thanks for the Foxie. Just love that little dog. But I think the following document which was the commencement of the program nails the theory I proposed about Brainchip, NASA, AWS and AKIDA technology with DARPA thrown in for good measure.
This paper was published by NASA with input from DARPA in May, 2019. Going way back in our searching we reached the point where it seemed clear that from perhaps August, 2019 NASA was aware of Brainchip and AKIDA technology.
Certainly in May, 2019 NASA was not officially aware when you read the entire document linked below not just the extract as they were relying upon three known neuromorphic technology developers all of which as has been proven were well behind AKIDA technology and as Peter van der Made has put in price sensitive announcements now do not tick all the boxes in particular do not have anything close to one shot and incremental learning. This feature as we know is ring fenced by those annoying things called patents owned by Brainchip.
Anyway I recommend reading the entire document as it makes clear the enormous reach of this project both from a space, cellestial and defence perspective. It also makes perfectly clear that the purpose of this project cannot be achieved without AKIDA technology and its one shot incremental learning. It also makes clear that this is a highly secretive technology advantage that NASA and DARPA are trying to achieve and I personally doubt that we will see any public announcement that AKIDA technology has been used to allow it to function for reasons which become obvious when you read the entire document.
The following is the extract that I thought of as having immediate interest. Shortly after this the document references the work that NASA had already done to try and achieve the AKIDA technology advantage with Brainchips failed competitors but as I said earlier the whole document should be read:
3.3.2.2 Cognitive Gateways
• Description: Cognitive gateways (CG) makes use of system and network requirements, observations, policies, and constraints in an effort to identify viable data paths and forward data objects using cognitive reasoning capabilities (Ref. 5). This effort was supported as part the Early Career Faculty (ECF) component of the Space Technology Research Grants Program.
• Capabilities Supported: Cognitive networking, e.g., adaptive data routing, “Drop data anywhere”, et al.
• Current Applications/Research: Adaptive Data Routing with Cognitive Gateways (Ref. 5)
○ Artificial learning using a modified reservoir of spiking neural networks
○ Spike firings are used to encode data network decisions
○ Evaluation of a prototype on an emulated space-earth testbed and through realistic simulations
• Potential Benefits:
○ Significant improvements to the reliability, performance, and asset efficiency of space
communication networks.
○ Low-power and small footprint implementation with cognitive processors.
○ Integrates multiple communication technologies: cognitive radio, delay-tolerant networks,
free-space optical, and conventional systems.
○ Advance core knowledge of cognitive networking applied to the context of NASA missions,
which involve harsh communication environments and the use of mixed network technologies.
○ Main innovation: New computational framework that exploits the brain-like computational power
of spiking neural network models for autonomous decision-making in space-earth data
networking
• Maturity Level: TRL 2/3
• Possible Infusion Points:
○ Ground stations and other terrestrial-based network nodes
○ Possible implementation on space-based assets using neuromorphic chips or other technologies
suitable for SWaP-constrained platforms
• Infusion Challenges: TRL advancement, e.g., implementation of routing capabilities on
SWaP-constrained platforms
This is the link:
https://ntrs.nasa.gov/api/citations/20190011723/downloads/20190011723.pdfMy two final comments would be that given all of the above and my earlier posts on this NASA technology link it would seem very clear that read together with all of the material unearthed by ManChild and uiux that Brainchip must in some form or other be engaged with Amazon Web Services even if it is only through NASA and DARPA. The second comment relates to AKD2000 and the implementation of LSTM predictive capabilities should not be discounted as the absolute cherry on the AKIDA technology cake as far as NASA and DARPA using Brainchip technology in this massive project.
As this is a very big call I state for the record this is my anonymous opinion only so DYOR - please and make sure you read it yourself and do not just take my word as I may have misread or misinterpreted something that materially changes your opinion.
FF
AKIDA BALLISTA