What I have noticed about Brainchip and AKIDA technology is that the more papers I read about edge/far edge/endpoint or a plethora of other used descriptors for this form of computing is that Brainchip and AKIDA technology solutions basically have been ignored by Academics.
The following two papers which are entirely devoted to an examination of edge computing the first of which was only released in September, 2021 and is even sponsored by the US Navy contain no reference to Brainchip and AKIDA technology.
These papers and multiple others all identify the absolute need for off loading much of the computing to smart end point devices rather than having the constant flow of data back and forth to the cloud be it public or private. There is no doubt that AKIDA technology is relevant to consider in such academic papers.In the first paper and the second the authors do not just deal with those technologies which they believe are suitable they deal with a range and advance reasons as to why one or other is found to be lacking in some way.
We as a group of shareholders hunger for details of the EAP and proof of concept customers that have not been released by Brainchip.Brainchip has consistently stated that due to Non Disclosure Agreements required by Brainchip and its prospective customers it cannot release details and that some of its prospective customers are alleged to have said that they will walk away if their first to market opportunity is undermined by exposure.
If you go looking for published academic articles by Peter van der Made they are actually relatively scarce when it comes to AKIDA technology however there are now 15 patents with his name assigned to Brainchip covering the AKIDA technology and more in the pipeline.This failure to produce academic papers is essential to ensure that his patents are granted and are effective.
The former CEO in 2020 when discussing the selection of EAP customers stated that they were only interested in real commercial opportunities and had no interest in university research projects and that their scarce resources would be entirely devoted to converting the selected commercial opportunities which would be selected from the well north of 100 NDA’s and would likely settle at about 2 dozen.The former CEO stated that among those selected would be household names and Fortune 500 companies.
So where is this all leading?
It is leading to the claim by Rob Lincourt DELL Technologies that they became interested in Brainchip and AKIDA technology because of some early figures Brainchip published regarding power consumption.How could it possibly be that Academics from around the world who constantly engage in literature searches completely missed these published figures, yet DELL Technologies saw them by chance and thought they were interesting enough to track down Brainchip and obtain a lend of the technology to explore its potential including deep diving into its capacity to be scaled for use as a coprocessor.
Brainchip is also the licensee of the JAST learning rules (STDP) and every second paper I pick up references these learning rules yet not one academic seems to have even by accident tripped over the AKIDA technology solution. @uiux even created a thread title “All Roads lead to JAST”.
Well I think the answer can be found if you take into account the following known facts:
1.The former CEO stated that they have no interest in University research projects,
2.Peter van der Made has maintained a very low profile academically in terms of publication which protects Brainchip’s patents; and
3.In such circumstances those who publish Academic papers must obtain written consent from Peter van der Made and others at Brainchip before information concerning AKIDA technology can be cited and used. It is a question of academic ethics.
Brainchip has in a calculated and deliberate fashion adopted a policy of ensuring limited if any academic exposure of the AKIDA technology solution.
Brainchip is a tiny player with revolutionary technology.Its greatest weakness was also its greatest strength. Being tiny the number of points of weakness regarding its intellectual property are reduced significantly. By remaining disengaged from the academic world, which as we know is inhabited by individuals who can be wined, dined, and bribed by multi nationals and foreign powers Brainchip has done everything possible to keep secure its intellectual property.
The risk of intellectual property threat is not hypothetical and in the USA in the very recent past there are examples of Professors at universities engaging in espionage and being arrested and convicted.
At the 2018 AGM the former CEO was asked about the prospects for a possible mobile telephone deal with a company in China and the former CEO specifically spoke to the risks around Brainchip’s intellectual property in that environment.In the end as we know Brainchip abandoned China for the time being mid-2020. Again, putting together his past comments it seems clear that protecting the intellectual property won out over the potential that China presented.
So where does this take us well it takes me at least to a point where it seems appropriate for Brainchip to strictly adhere to the current EAP and NDA agreements and to ignore the calls from shareholders for more information.
The share price is secondary to this need to protect Brainchip’s intellectual property and customers secrecy requirements.In my opinion all shareholders including myself should allow Brainchip and its trusted advisors to continue with their strategic commercial plan aimed at protecting the intellectual property of the company and thereby ensuring its future commercial success.
The present reality is that with the testing of the commercial AKD1000 and the proposed release thereof to small engineering firms, EAP’s and proof of concept customers as well as the Tacheon Pi and other modules Brainchip is heralding the next phase in its pathway to commercial success and so unlike in 2016 when I first started to research Brainchip the endpoint is incredibly close in comparison.
My opinion only DYOR
FF
AKIDA BALLISTA
A System Design Perspective on Neuromorphic ComputerProcessors
Garrett S. Rose1 , Mst Shamim Ara Shawkat2 , Adam Z. Foshie1 , John J. Murray VI1 , and Md Musabbir Adnan1 1Min H. Kao Department of Electrical Engineering & Computer Science, The University of Tennessee, Knoxville, TN 37996, USA 2Department of Electrical and Computer Engineering, Florida International University, Miami, FL 33174, USA E-mail: [email protected] MONTH 2021
Abstract.
Neuromorphic computing has become an attractive candidate for emerging computing platforms. It requires an architectural perspective, meaning the topology or hyperparameters of a neural network is key to realizing sound accuracy and performance in neural networks. However, these network architectures must be executed on some form of computer processor. For machine learning, this is often done with conventional computer processing units, graphics processor units, or some combination thereof. A neuromorphic computer processor or neuroprocessor, in the context of this paper, is a hardware system that has been designed and optimized for executing neural networks of one flavor or another. Here, we review the history of neuromorphic computing and consider various spiking neuroprocessor designs that have emerged over the years. The aim of this paper is to identify emerging trends and techniques in the design of such brain-inspired neuroprocessor computer systems.
Acknowledgments
This material is based in part on research sponsored by Air Force Research Laboratory under agreement number FA8750-19-1-0025. The U.S. Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of Air Force Research Laboratory or the U.S. Government
https://iopscience.iop.org/article/10.1088/2634-4386/ac24f5/pdf
Edge Machine Learning forAI-Enabled IoT Devices: A Review
Abstract and Figures
Ina few years, the world will be populated by billions of connected devices thatwill be placed in our homes, cities, vehicles, and industries. Devices withlimited resources will interact with the surrounding environment and users.Many of these devices will be based on machine learning models to decodemeaning and behavior behind sensors’ data, to implement accurate predictionsand make decisions. The bottleneck will be the high level of connected thingsthat could congest the network. Hence, the need to incorporate intelligence onend devices using machine learning algorithms. Deploying machine learning onsuch edge devices improves the network congestion by allowing computations tobe performed close to the data sources. The aim of this work is to provide areview of the main techniques that guarantee the execution of machine learningmodels on hardware with low performances in the Internet of Things paradigm,paving the way to the Internet of Conscious Things. In this work, a detailedreview on models, architecture, and requirements on solutions that implementedge machine learning on Internet of Things devices is presented, with the maingoal to define the state of the art and envisioning development requirements.Furthermore, an example of edge machine learning implementation on amicrocontroller will be provided, commonly regarded as the machine learning“Hello World”.
https://www.researchgate.net/publication/341037496_Edge_Machine_Learning_for_AI-Enabled_IoT_Devices_A_Review
(20min delay)