Bravo
Meow Meow 🐾
Not exactly. But that made me think of another one.
Question: What do you get if you cross BrainChip with an un-ironed tablecloth?
Answer: The Wrinkle-Gate Affair.
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Not exactly. But that made me think of another one.
Question: What do you get if you cross BrainChip with an un-ironed tablecloth?
Answer: The Wrinkle-Gate Affair.
Frank Zappa
Regular
Looks like it was not something that she could do all that often
Masterofdesaster
Emerged
ADA100 or AKIDA?
www.eenewsanalog.com
Neuromorphic startup Reexen gains ground with Goertek
Analog and mixed-signal neuromorphic startup Reexen Technology Co. Ltd. has achieved a key design win with microphone assembler Goertek.
www.eenewsanalog.com
VictorG
Member
You have killed several billion of my synapse.
JK200SX
Regular
AI Powered Hearing Aids | Starkey Livio Edge AI
Starkey Livio Edge AI is the world’s first custom rechargeable hearing aid to deliver next-generation sound performance. Our technology, Edge Mode, instantaneously adjusts to help you hear in the most challenging listening situations.www.starkey.com.au
Potential.....
Neuromorphia
fact collector
Context...Brainchip job advertisement mentions Agricultural Equipment. Here is a video with autonomous Agricultural Equipment.
Last edited:
Fact Finder
Top 20
In one of Mr. Dinardo the former CEO’s webinars he spoke in an almost random fashion about the fact that he had no idea how many different data points tractors measured and sent to the cloud. He did put a number which I think was over 100.
There has been clearly an ongoing engagement with the Agricultural industry for some years. We have even had photographs of tractors in presentations.
We know at the highest level of probability that an EAP is one of the large players in advanced intelligent automated agriculture we just don’t know which one.
My opinion only DYOR
FF
AKIDA BALLISTA
Easytiger
Regular
WowTaken from the MegaChips website, the text below shows that BrainChip is an important part of MegaChip's business model in the US. This is a strategic move made by a successful Japanese company. I believe a great deal of research would have been conducted by MegaChips prior to them selecting BrainChip as one of their two AI IP providers.
Moving to the US
MegaChips came to the United States in 1995, but kept a low profile. However it has made a number of strategic investments in the US. It acquired SiTime Corp, located in Santa Clara CA, in 2014 and then spun it out as a public company in 2019. In July 2021, MegaChips invested in SiliconBrite, a company focused on analog and mixed-signal technologies. Later in 2021, MegaChips struck a strategic partnership with Motus-Labs to work jointly on products for the robotics and automated equipment.
In a much more aggressive move, in mid-2020, founder and Chairman, Masahiro Shindo, identified AI/ML technology to be critical to Megachips’ future and asked the US operation to take a leadership position in moving the company in that direction.
MegaChips began an internal training program to allow a group of dedicated engineers to become experts in this important technology. The company made significant investments in the US to identify key partners, build relationships with local universities, and acquire key talent in this space. In 2021, the company made multi-million-dollar investments in two key AI/IP partners, Brainchip and Quadric, to bolster its offerings in the Edge AI market. The company is now positioned to make an aggressive move into the US ASIC market, using its skills in Edge AI as a key component of that move.
View attachment 3432
Masahiro Shindo
About MegaChips | MegaChips Corporation
Founded in 1990 as Japan's first manufacturer of R&D-oriented system LSIs, MegaChips leads the world market through innovation and creativity.megachips.com
Cyw
Regular
Wonder where those multi-million dollar investments went?
Hi FF,
I noticed you were looking askance at my linking NASA and MemRistors.
If you google "NASA memristor" you will find a wealth of SBIRs, over several years, eg:
https://sbir.nasa.gov/SBIR/abstracts/21/sbir/phase1/SBIR-21-1-H6.22-2890.html
Radiation-Hardened Memristor-based Memory for Extreme Environments
Submitted by drupal on Wed, 10/23/2013 - 18:04Early investigations have also shown that memristors have high radiation hardness. In this SBIR, CFDRC and Arizona State University propose to develop, characterize, and demonstrate novel, memristor-based, radiation-hardened NVM (non-volatile memory) for NASA space applications. In Phase I we will: 1) Fabricate state-of-the-art Chalcogenide Glass (ChG) memristors based on the CBRAM technology; 2) Examine their wide temperature performance (-230 to +130 deg.C) via thermal experiments; and 3) Add new models to CFDRC's NanoTCAD Mixed-Mode simulator for accurate physics-based simulation of memristors. The Phase I effort will evaluate suitability of ChG memristors for extreme temperature applications. In Phase II, we will extend our scope to include wide-temperature investigation of the competing transition-metal-oxide (TMO, e.g., TiO2) memristor technology. For both ChG and TMO, we will then perform irradiation testing and down-select the technology with the best extreme environment (radiation + temperature) performance. Subsequently, we will generate wide-temperature, radiation-enabled, device physics and compact models for the memristors, develop designs for memristor-based NVM, and perform mixed-mode simulations to determine their radiation and thermal response. These results, and physics-based understanding of device response, will be used to develop an NVM prototype that will be tested and demonstrated for NASA space applications.
#######################################################
https://sbir.nasa.gov/SBIR/abstracts/12/sbir/phase1/SBIR-12-1-H6.02-9737.html
Proposal Number: 21-1- H6.22-2890
Subtopic Title:
Deep Neural Net and Neuromorphic Processors for In-Space Autonomy and Cognition
Proposal Title:
3D Integrated Memristor Chip for Neuromorphic Processing
Form Generated on 04/06/2021 12:12:34
The innovation is in the design and fabrication of oxide thin film based memristors for each layer, with some of the memristors exhibiting large off/on resistance ratios, some used for switching, and some with low off/on resistance ratio with large number of intermediate resistance states (depicting the synaptic weights). Innovative fabrication and packaging solutions will be considered in Phase 2 for developing a microsystem with unique low power neuromorphic computing capability with SWaP considerations for small satellites.
Potential NASA Applications (Limit 1500 characters, approximately 150 words):
Neuromorphic computing is expected to enable NASA’s growing demands for artificial intelligence (AI) and machine learning (ML) on board space platforms to optimize and automate operations. A hardware based neuromorphic computing utilizing memristors provides a potential low power computing with integrated memory and decentralized operations. Such an architecture could be used for onboard learning to optimize communication and data processing capabilities in a cognitive system meeting the SWaP constraints in space systems.
While it may be that there is scope for the use of memristor NVM in conjunction with Akida, memristors have a tendency for the charge to leak, which is a serious problem for analog NNs. Only the other day we looked at a hybrid (Frankenstein) NN (Syntiant) which used memristors, but required a digital output stage to correct for the drift in memristor charge.
BRN Discussion 2022
Hi Vanman, We've looked at Syntiant a couple of times here and "there". They have a Frankenstein neural arrangement blending analog and digital circuitry. We looked at Syntiant earlier this (last) year, but they are worth another look. Many of their patents relate to the computerized...- Diogenese
- Post #6,730
- Yesterday at 8:27 PM
- Stock: BRN - Brainchip Holdings Ltd
"3. Why could AKIDA technology not be providing the extra processing power to maximise the analogue graphic processing unit???
“graphic processor units that use electric analog circuits to mimic the human nervous system, known as or neuromorphic processors, adds much more "thinking" capability in these small boxes”"
I think it is an either/or situation, they would both be doing the same job of being NNs classifying input data. So, we know that NASA also has a thing for redundancy, so they could have one of each. Better still, they could have 3 Akidas.
Last edited:
alwaysgreen
Top 20
Oh Deere. I hope whichever company it is, we sell them on Masse-y.
S
Straw
Guest
As long as they don't retrofit fully AGI capable vibration analysis on 60yo acreage tractors or it might develop anxiety issues.
Terroni2105
Founding Member
In person Poster Presentation ML summit Tuesday 29 March 2022 ( poster below)
Do what you need to over my head DYOR
Built around a mesh-connected array of neural processor units (NPUs) the architecture is highly scalable to meet the needs of a wide range of applications. The uniqueness of the BrainChip Akida Architecture lies in the ability of the hardware to run traditional feedforward, deeply learned CNN networks as well as native SNN networks. This documentation provides examples of how to develop both classes of solutions, using industry standard tool flows and networks, to solve a variety of application problems such as vision, acoustic, cybersecurity amongst others.
The Akida neural processor is available both as Intellectual Property (IP) circuit design for integration in ASIC products or as a System on a Chip (SoC) product.
As Figure 1 shows, the SoC is built around a core neural processor comprised of 80 neural processing units, it includes a conversion complex and allows one to run popular convolutional neural networks (CNNs) such as MobileNet 1. Designers can use the Akida SoC to run industry standard CNNs, dramatically reducing power by changing convolutions to event based computations, or run native SNN solutions.
Figure 1. BrainChip Akida processor
The Akida chip includes several key features that differentiate it from other neural network processors and deep learning accelerators. These are:
Figure 2. Akida IP example configurations
The core data structure used by the Akida runtime is a neural network model, which itself is a linear stack of layers.
MetaTF leverages the TensorFlow framework and PyPI for BrainChip tools installation. The major difference with other machine learning frameworks is that the data exchanged between layers is not the usual dense multidimensional arrays, but sets of spatially organized events that can be modelled as sparse multidimensional arrays.
Throughout this documentation, those events will often be referred as “spikes”, due to their close similarity with the signals exchanged by biological neurons.
Figure 3. Akida MetaTF ML Framework
The MetaTF ML framework comprises three main python packages:
Figure 4. Akida python package
Figure 5. Akida runtime configurations
Note
While the Akida examples are provided under an Apache License 2.0, the underlying Akida library is proprietary.
Please refer to the End User License Agreement for terms and conditions.
1
In most cases the entire network can be accommodated using the on-chip SRAM. Even the large MobileNet network used to classify 1000 classes of ImageNet fits comfortably.
Next
© Copyright 2022, BrainChip Holdings Ltd. All Rights Reserved.
Do what you need to over my head DYOR
listed document in the poster
Overview — Akida Examples documentation
doc.brainchipinc.com
Overview — Akida Examples documentation
Overview
The Akida Neural Processor
BrainChip’s Akida integrated circuit technology is an ultra-low power, high performance, minimum memory footprint, event domain neural processor targeting Edge AI applications. In addition, because the architecture is based upon an event domain processor, leveraging fundamental principles from biological SNNs, the processor supports incremental learning. This allows a deeply trained network to continue to learn new classifiers without requiring a re-training process. Due to the highly optimized architecture, the Akida Neural Processor eliminates the need for a CPU to run the neural network algorithm and in most cases eliminates the need for a DRAM external to the neural fabric. The elimination of external devices makes the Akida solution significantly more power efficient compared to deep learning accelerators which require both external CPU and memory.Built around a mesh-connected array of neural processor units (NPUs) the architecture is highly scalable to meet the needs of a wide range of applications. The uniqueness of the BrainChip Akida Architecture lies in the ability of the hardware to run traditional feedforward, deeply learned CNN networks as well as native SNN networks. This documentation provides examples of how to develop both classes of solutions, using industry standard tool flows and networks, to solve a variety of application problems such as vision, acoustic, cybersecurity amongst others.
The Akida neural processor is available both as Intellectual Property (IP) circuit design for integration in ASIC products or as a System on a Chip (SoC) product.
As Figure 1 shows, the SoC is built around a core neural processor comprised of 80 neural processing units, it includes a conversion complex and allows one to run popular convolutional neural networks (CNNs) such as MobileNet 1. Designers can use the Akida SoC to run industry standard CNNs, dramatically reducing power by changing convolutions to event based computations, or run native SNN solutions.
Figure 1. BrainChip Akida processor
The Akida chip includes several key features that differentiate it from other neural network processors and deep learning accelerators. These are:
- Event-based computing leveraging inherent data and activation sparsity
- Fully configurable neural processing cores, supporting convolutional, separable-convolutional, pooling and fully connected layers
- Incremental learning after off-line training
- On-chip few-shot training
- Configurable number of NPUs
- Programmable data to event converter
- Event-based NPU engines running on a single clock
- Configurable on-chip SRAM memory
- Runs full neural networks in hardware
- On chip communication via mesh network
- On chip learning in event domain
- Process technology independent platform
- Network size customizable to application needs
- IP deliverables include: RTL, dev tools, test suites and documentation
Figure 2. Akida IP example configurations
The Akida Neuromorphic ML Framework
The Akida Neuromorphic ML Framework (MetaTF) relies on a high-level neural networks API, written in Python, and largely inspired by the Keras API.The core data structure used by the Akida runtime is a neural network model, which itself is a linear stack of layers.
MetaTF leverages the TensorFlow framework and PyPI for BrainChip tools installation. The major difference with other machine learning frameworks is that the data exchanged between layers is not the usual dense multidimensional arrays, but sets of spatially organized events that can be modelled as sparse multidimensional arrays.
Throughout this documentation, those events will often be referred as “spikes”, due to their close similarity with the signals exchanged by biological neurons.
Figure 3. Akida MetaTF ML Framework
The MetaTF ML framework comprises three main python packages:
- the Akida python package is an interface to the Brainchip Akida Neuromorphic System-on-Chip (NSoC). To allow the development of Akida models without an actual Akida hardware, it includes a runtime, an Hardware Abstraction Layer (HAL) and a software backend that simulates the Akida NSoC (see Figure 4 and Figure 5).
- the CNN2SNN tool provides means to convert Convolutional Neural Networks (CNN) that were trained using Deep Learning methods to event domain, low-latency and low-power network for use with the Akida runtime.
- the Akida model zoo contains pre-created network models built with the Akida sequential API and the CNN2SNN tool using quantized Keras models.
Figure 4. Akida python package
Figure 5. Akida runtime configurations
The Akida examples
The examples section comprises a zoo of event-based CNN and SNN tutorials. One can check models performances against MNIST, ImageNet and Google Speech Commands (KWS) datasets.Note
While the Akida examples are provided under an Apache License 2.0, the underlying Akida library is proprietary.
Please refer to the End User License Agreement for terms and conditions.
1
In most cases the entire network can be accommodated using the on-chip SRAM. Even the large MobileNet network used to classify 1000 classes of ImageNet fits comfortably.
Next
© Copyright 2022, BrainChip Holdings Ltd. All Rights Reserved.
Last edited:
TechGirl
Founding Member
Good to see this ISL story is still getting around
https://inf.news/en/science/791bf988657af0e134c475ca748067c6.html
BrainChip selected by U.S. Air Force Research Laboratory to develop AI-based radar
2022-03-30 22:12 HKT
BrainChip, the world's first commercial producer of neuromorphic artificial intelligence chips and IP, today announced that Information Systems Laboratories (ISL) is developing an AI-based radar for the U.S. Air Force Research Laboratory (AFRL) based on its Akida™ Neural Network Processor Research solutions.
ISL is a specialist in expert research and complex analysis, software and systems engineering, advanced hardware design and development, and high-quality manufacturing for a variety of clients worldwide.
ISL focuses on areas such as advanced signal processing, space exploration, subsea technology, surveillance and tracking, cybersecurity, advanced radar systems and energy independence. As a member of the BrainChip Early Partnership Program (EAP), ISL will be able to evaluate boards for Akida devices, software and hardware support, and dedicated engineering resources.
"As part of BrainChip's EAP, we had the opportunity to directly assess the capabilities Akida offers to the AI ecosystem," said Jamie Bergin, Senior Vice President, Research, Development and Engineering Solutions Manager at ISL.
BrainChip brings AI to the edge in ways not possible with existing technologies. Akida processors feature ultra-low power consumption and high performance to support the development of edge AI technologies by using neuromorphic architecture, a type of artificial intelligence inspired by the biology of the human brain. BrainChip's EAP program provides partners with the ability to realize significant benefits of power consumption, design flexibility and true learning at the edge.
"ISL's decision to use Akida and Edge-based learning as a tool to incorporate into their research and engineering solutions portfolio is in large part due to the go-to-market advantages our innovation capabilities and production-ready status provide ” said Sean Hehir, CEO of BrainChip, “We are delighted to be a partner of AFRL and ISL on edge AI and machine learning. We believe the combination of technologies will help accelerate the deployment of AI in the field.”
Akida is currently licensed as IP and is also available to order for chip production. It focuses on low power consumption and high performance, supports sensory processing, and is suitable for applications that benefit artificial intelligence, as well as applications such as smart healthcare, smart cities, smart transportation, and smart homes.
https://inf.news/en/science/791bf988657af0e134c475ca748067c6.html
BrainChip selected by U.S. Air Force Research Laboratory to develop AI-based radar
2022-03-30 22:12 HKT
BrainChip, the world's first commercial producer of neuromorphic artificial intelligence chips and IP, today announced that Information Systems Laboratories (ISL) is developing an AI-based radar for the U.S. Air Force Research Laboratory (AFRL) based on its Akida™ Neural Network Processor Research solutions.
ISL is a specialist in expert research and complex analysis, software and systems engineering, advanced hardware design and development, and high-quality manufacturing for a variety of clients worldwide.
ISL focuses on areas such as advanced signal processing, space exploration, subsea technology, surveillance and tracking, cybersecurity, advanced radar systems and energy independence. As a member of the BrainChip Early Partnership Program (EAP), ISL will be able to evaluate boards for Akida devices, software and hardware support, and dedicated engineering resources.
"As part of BrainChip's EAP, we had the opportunity to directly assess the capabilities Akida offers to the AI ecosystem," said Jamie Bergin, Senior Vice President, Research, Development and Engineering Solutions Manager at ISL.
BrainChip brings AI to the edge in ways not possible with existing technologies. Akida processors feature ultra-low power consumption and high performance to support the development of edge AI technologies by using neuromorphic architecture, a type of artificial intelligence inspired by the biology of the human brain. BrainChip's EAP program provides partners with the ability to realize significant benefits of power consumption, design flexibility and true learning at the edge.
"ISL's decision to use Akida and Edge-based learning as a tool to incorporate into their research and engineering solutions portfolio is in large part due to the go-to-market advantages our innovation capabilities and production-ready status provide ” said Sean Hehir, CEO of BrainChip, “We are delighted to be a partner of AFRL and ISL on edge AI and machine learning. We believe the combination of technologies will help accelerate the deployment of AI in the field.”
Akida is currently licensed as IP and is also available to order for chip production. It focuses on low power consumption and high performance, supports sensory processing, and is suitable for applications that benefit artificial intelligence, as well as applications such as smart healthcare, smart cities, smart transportation, and smart homes.
Ok, this next BrainChip podcast with Ian Drew intrigues me. This is the chairman of Foundries.io, a company that partners with other tech companies to provide edge solutions. He is speaking with BrainChip about next generation IOT technology. Akida is next generation IOT technology. How do you get through a podcast like this without speaking about Akida???? And why hasn't Foundries.io already partnered with BrainChip as they have with the companies listed below? This will be interesting!!!
Krustor
Regular
Found by user Woody_8 from the German BRN-community:
A new graphic illustration from a car. Definitifely looks like a Dodge Charger. Especially the illustration of the doors.
Dodge is owned by Stellantis, the fourth biggest car manufacturer worldwide. This would be huge.
A new graphic illustration from a car. Definitifely looks like a Dodge Charger. Especially the illustration of the doors.
Dodge is owned by Stellantis, the fourth biggest car manufacturer worldwide. This would be huge.
Krustor
Regular
Seems like it is just a random stock picture: https://stock.adobe.com/de/images/f...tor-illustration/294492535?asset_id=320902801
Fact Finder
Top 20
A simple to understand explanation as to why AKIDA cortical columns will blow the semiconductors market socks off.
My opinion only DYOR
FF
AKIDA BALLISTA
Applied Physics Reviews 9, 011313 (2022);https://doi.org/10.1063/5.0062603
B. A. W. Brinkman1,2,
H. Yan3,
A. Maffei1,2,
I. M. Park1,2,
A. Fontanini1,2,
J. Wang2,4,a), and
G. La Camera1,2,a)
View AffiliationsView Contributors
ABSTRACT
Cortical neurons emit seemingly erratic trains of action potentials or “spikes,” and neural network dynamics emerge from the coordinated spiking activity within neural circuits. These rich dynamics manifest themselves in a variety of patterns, which emerge spontaneously or in response to incoming activity produced by sensory inputs. In this Review, we focus on neural dynamics that is best understood as a sequence of repeated activations of a number of discrete hidden states. These transiently occupied states are termed “metastable” and have been linked to important sensory and cognitive functions. In the rodent gustatory cortex, for instance, metastable dynamics have been associated with stimulus coding, with states of expectation, and with decision making. In frontal, parietal, and motor areas of macaques, metastable activity has been related to behavioral performance, choice behavior, task difficulty, and attention. In this article, we review the experimental evidence for neural metastable dynamics together with theoretical approaches to the study of metastable activity in neural circuits. These approaches include (i) a theoretical framework based on non-equilibrium statistical physics for network dynamics; (ii) statistical approaches to extract information about metastable states from a variety of neural signals; and (iii) recent neural network approaches, informed by experimental results, to model the emergence of metastable dynamics. By discussing these topics, we aim to provide a cohesive view of how transitions between different states of activity may provide the neural underpinnings for essential functions such as perception, memory, expectation, or decision making, and more generally, how the study of metastable neural activity may advance our understanding of neural circuit function in health and disease.
My opinion only DYOR
FF
AKIDA BALLISTA
- Metastable dynamics of neural circuits and networks
-
Applied Physics Reviews 9, 011313 (2022);https://doi.org/10.1063/5.0062603
B. A. W. Brinkman1,2,
H. Yan3,
A. Maffei1,2,
I. M. Park1,2,
A. Fontanini1,2,
J. Wang2,4,a), and
G. La Camera1,2,a)View AffiliationsView Contributors
ABSTRACT
Cortical neurons emit seemingly erratic trains of action potentials or “spikes,” and neural network dynamics emerge from the coordinated spiking activity within neural circuits. These rich dynamics manifest themselves in a variety of patterns, which emerge spontaneously or in response to incoming activity produced by sensory inputs. In this Review, we focus on neural dynamics that is best understood as a sequence of repeated activations of a number of discrete hidden states. These transiently occupied states are termed “metastable” and have been linked to important sensory and cognitive functions. In the rodent gustatory cortex, for instance, metastable dynamics have been associated with stimulus coding, with states of expectation, and with decision making. In frontal, parietal, and motor areas of macaques, metastable activity has been related to behavioral performance, choice behavior, task difficulty, and attention. In this article, we review the experimental evidence for neural metastable dynamics together with theoretical approaches to the study of metastable activity in neural circuits. These approaches include (i) a theoretical framework based on non-equilibrium statistical physics for network dynamics; (ii) statistical approaches to extract information about metastable states from a variety of neural signals; and (iii) recent neural network approaches, informed by experimental results, to model the emergence of metastable dynamics. By discussing these topics, we aim to provide a cohesive view of how transitions between different states of activity may provide the neural underpinnings for essential functions such as perception, memory, expectation, or decision making, and more generally, how the study of metastable neural activity may advance our understanding of neural circuit function in health and disease.
This could be the telecommunications dot that was being discussed, speculation only but Rob seems to like it.
.
.
TEAL Hits $100M Valuation, Citing 1,032% Recurring Revenue Growth in 10-Month Period
Seattle-based startup TEAL announced $10.8 million in new funding. TEAL was founded by Robert Hamblet and Michael Johnston to bring programmable networking services to the Internet of Things (IoT) industry.
Attachments
Similar threads
