Getupthere
Regular
BrainChip Confirms Validation of the AkidaTM Neural Processor
14th of September 2020 and the EAP clients had access to akida well before this.
RENESAS signed Ip license on the 23rd of December 2020.
Last edited:
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
14th of September 2020 and the EAP clients had access to akida well before this.
RENESAS signed Ip license on the 23rd of December 2020.
Emerging+Neuromorphic+Computing+for+Edge+AI+Application+(1).pdf
TITLE:
"Emerging Neuromorphic Computing for Edge AI Application: A Systematic Literature Review"
AKIDA gets a mention:
· "Akida [23] is considered the first commercial neuromorphic processor. Akida is advertised as a power-efficient event-based processor for edge
computing, which does not require any external CPU. Akida is capable of processing at 1000 watts. Around 80 Neural processing units are used in 1 Akida chip in a mesh network, which enables 1,200,000 neurons and 10,000,000,000 synapses. Akida chip was primarily built at TSMC nm. Further, a free chip emulator is also provided using the Akida ecosystem."
Merits of Neuromorphic listed as - i have summarised below, the report contains an explanation of each of them.:
Lower power consumption, Robustness and fault tolerance, Sensor fusion and perception, Reduced latency, · Reduced Bandwidth and storage requirement, Proficient resource utilization.
TITLE:
"Emerging Neuromorphic Computing for Edge AI Application: A Systematic Literature Review"
AKIDA gets a mention:
· "Akida [23] is considered the first commercial neuromorphic processor. Akida is advertised as a power-efficient event-based processor for edge
computing, which does not require any external CPU. Akida is capable of processing at 1000 watts. Around 80 Neural processing units are used in 1 Akida chip in a mesh network, which enables 1,200,000 neurons and 10,000,000,000 synapses. Akida chip was primarily built at TSMC nm. Further, a free chip emulator is also provided using the Akida ecosystem."
Merits of Neuromorphic listed as - i have summarised below, the report contains an explanation of each of them.:
Lower power consumption, Robustness and fault tolerance, Sensor fusion and perception, Reduced latency, · Reduced Bandwidth and storage requirement, Proficient resource utilization.
BRN's chief advantage is Edge AI. We are already there with Edge Box/Server, M.2 PCB, Frontgrade space, Onsor,... and then there are the NDAs.
I think one dot that's missing is Cybersecurity AI - that really is essential AI ... and we're there now, just on the edge of the plateau of productivity.
Don't know why they have neuromorphic computing at 5+ years - done and dusted.
Then there's GenAI, due 2026 in the Roadmap. I'm guessing we will have RAG (Retrieval Augmented Generation) GenAI which is specialized by topic/model.
Where the hype cycle graph is wrong is that, for AI, the plateau of productivity will be miles higher than the Peak of Inflated Expectations, more of a hockey stick.
Bravo
If ARM was an arm, BRN would be its biceps💪!
OK. So you're that guy on the Crapper that keeps whinging and wining over and over and over again?
Is that you?
And you chose to pick on me earlier? Wow! Tough guy!
Hehehe!
equanimous
Norse clairvoyant shapeshifter goddess
AKIDA1000 was released in 2020 but the commercial push was not started until late 2021 when Vianna (NED and Commercial and Operational consultant) and Hehir were appointed.
Jerome Nadel was also appointed Chief Marketing Officer in Jan'22.
Getupthere
Regular
You do realise that 12 EAP had access to Akida 1 before the Akd1000 was released, well before 2020.
Sean’s five-year plan has been to create ecosystems and for young students in partnering universities who can access Akida.
While he was doing this, his five-year plan was to sign new IP deals, but this hasn’t worked out as planned.
Last edited:
White Horse
Regular
Fuck off. !!
Similar threads
