aTENNuate for raw speech denoising in real time, page-20

I was having a look over somepast research papers and came across the Quantum Ventura work from 2022. It hasbeen discussed previously but the following paragraph jumped out at me:

“We sought to evaluate theviability of a real-time HPC-scale neuromorphic cybersecurity system calledCyber- Neuro RT. We began by developing a full precision deep learning (DL)network for 450,000 Zeek log entries with a mixture of normal and maliciousdata, including eight different types of attacks. Deep learning to spikingneural network (DL-to-SNN) conversions were conducted for two vendors, IntelLoihi and BrainChip AkidaTM to explore their platforms. Design spaceexplorations to increase SNN inference speed were also conducted. Neuromorphicimplementations of deep learning networks frequently achieve the same level ofaccuracy as full precision models while saving substantial power and cost. Weinvestigate this statement in the context of cybersecurity. The results arepromising, but more research is needed.”

https://www.sciencedirect.com/science/article/pii/S1877050922017938/pdf?md5=d9c849d37eb0815269240514a9efbe22&pid=1-s2.0-S1877050922017938-main.pdf

The above paper was clearly partof the research effort by Quantum Ventura to bring a neuromorphic ‘CyberNeuroRT’ solution to market. The fact that Quantum Ventura has now know releasedcommercial CyberNeuro RT product offerings for both Loihi and AKD1000 confirmsthat the further research had a positive outcome regarding the use of AKIDAtechnology for cyber security. This isan irrefutable logical conclusion as we have the tangible proof in the belowlink that one can actually purchase from Quantum Ventura a CyberNeuro RT AKIDAtechnology powered solution today.

https://www.quantumventura.tech/cyberneuro-rt

This brings me to thespeculative part of this post. Looking at the above evidence it seems highlylikely that the SBIR Phase 2 grant to Quantum Ventura and Brainchip fordesigning a cyber security system to protect infrastructure funded by the USDept of Energy has had, or will likely have, a happy ending.

I believe with a high level ofconfidence however that if my speculation is correct then given the need tomaintain secrecy around how the US Dept of Energy protects essential infrastructurefrom bad actors, I will not expect confirmation any time soon. I do note however what Quantum Venturastated:

Quantum Ventura, a San Jose-based provider of AI and ML research andtechnologies, is using BrainChip’s Akida technology to develop cyberthreat-detection tools.

Australianfirm BrainChip is a commercial producer of low power, digital, event-based,neuromorphic AI IP.

Ina federally funded phase-two programme, Quantum Ventura is creatingcyber-security applications for the US Department of Energy under the SmallBusiness Innovation Research (SBIR) programme. The programme is focused oncyber threat-detection using neuromorphic computing, which aims to develop anapproach to detect and prevent cyber attacks on computer networks and criticalinfrastructure using brain-inspired artificial intelligence.

“Neuromorphiccomputing is an ideal technology for threat detection because of its small sizeand power, accuracy, and, in particular, its ability to learn and adapt, sinceattackers are constantly changing their tactics,” said Srini Vasan, CEO ofQuantum Ventura. “We believe that our solution incorporating BrainChip’s Akidawill be a breakthrough for defending against cyber threats and addressadditional applications as well.”https://www.iotm2mcouncil.org/iot-library/news/iot-newsdesk/brainchip-ai-detects-cyber-threats-for-us-government/

The following is the link to theoriginal announcement by Brainchip and Quantum Ventura:

https://brainchip.com/brainchip-and-quantum-ventura-partner-to-develop-cyber-threat-detection/

My opinion only DYOR

Fact Finder