BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Falling Knife
Regular
Chat GPT is old data from 2021
Not a bad list from which to grow?
Which of these or others will be involved in the current discussed launch?
Fact Finder
Top 20
Interesting how Nintendo is the only customer they admit to having and in 2021 they accounted for 75% of revenue if I recall that number correctly and ChatGpt did not include it in the list. Also interesting that Apple which states it is a customer of MegaChips on its own website does not appear. It does have Sony which I found early on when MegaChips first was announced.
My opinion only DYOR
FF
AKIDA BALLISTA
chapman89
Founding Member
From the Megachips “Business & Other Risks section as of 2022-
Forward-looking statements in this section represent the judgment of MegaChips as of March 31, 2022.
for OA equipment. The percentage of net sales involving LSI for storing game software (custom memory) to Nintendo Co., Ltd. (“Nintendo”) is increasing and accounts for 87.8% of the sales for the current fiscal year.
Therefore, the performance of the Company could fluctuate depending on the sales trend of the game consoles and software that use our LSI products, and the market of LSI.
The risk is not something that can be completely eliminated, we have the good and close relationship with Nintendo and aim to meet customer satisfaction with our products by providing ideal solution and stable supply and minimize the possible risks. Besides, we focus on the development of the new business in the fields including industrial equipment, telecommunications, AI, energy control and robotics as well as improving the business portfolio in the mid- to long-term”
“Currently MegaChips Group focuses its management resources in the growth areas such as in-vehicle device, industrial equipment, telecommunication infrastructure, energy control and robotics and working on the research and development to provide state-of-the-art technology and products to the customers. The R&D expenditures totaled ¥2,537 million and accounted for 3. 4% of consolidated sales”
Forward-looking statements in this section represent the judgment of MegaChips as of March 31, 2022.
Dependence on specific customers
Purchasers
MegaChips principally sells LSI for a game software storage (custom memory) for the amusement field; LSI for game consoles and peripheral devices; LSI for digital cameras and other image processing; and LSI (Limited Social Interaction)for OA equipment. The percentage of net sales involving LSI for storing game software (custom memory) to Nintendo Co., Ltd. (“Nintendo”) is increasing and accounts for 87.8% of the sales for the current fiscal year.
Therefore, the performance of the Company could fluctuate depending on the sales trend of the game consoles and software that use our LSI products, and the market of LSI.
The risk is not something that can be completely eliminated, we have the good and close relationship with Nintendo and aim to meet customer satisfaction with our products by providing ideal solution and stable supply and minimize the possible risks. Besides, we focus on the development of the new business in the fields including industrial equipment, telecommunications, AI, energy control and robotics as well as improving the business portfolio in the mid- to long-term”
“Currently MegaChips Group focuses its management resources in the growth areas such as in-vehicle device, industrial equipment, telecommunication infrastructure, energy control and robotics and working on the research and development to provide state-of-the-art technology and products to the customers. The R&D expenditures totaled ¥2,537 million and accounted for 3. 4% of consolidated sales”
Business and Other Risks | MegaChips Corporation
We aim at sustainable growth through profitability and growth opportunity.
www.megachips.co.jp
Last edited:
Falling Knife
Regular
Yes it only has the grades of a just above average University graduate and the data set has not been updated since 2021.Well chat GPT is a bit of a loser at times since no mention of Nintendo
Snap Fact Finder'beat me to it.
Even removing Nintendo we have a list the BRn SP would spring off very nicely given half a chance mention in an ASX announcement
that has $ attached of course Rise from the ashes
Regular
Well it's smarter than me thenYes it only has the grades of a just above average University graduate and the data set has not been updated since 2021.
Even removing Nintendo we have a list the BRn SP would spring off very nicely given half a chance mention in an ASX announcement
Yes that day will come soon hopefully none of us die before then
Steve10
Regular
An article written by Douglas Fairbairn from MegaChips.
As the need for artificial intelligence grows more common and technology needs become more sophisticated, companies looking to adopt edge AI into their products often find it to be a difficult challenge. But what makes it so difficult, and what solutions exist to solve this problem?
Perhaps the single biggest issue that companies face in implementing edge AI is that most companies don’t have the resources in house to develop these sophisticated fast-changing technologies. Lack of trained personnel and little familiarity with design flow often leads to delayed timelines and excess expense to train team members. In addition, there are so many choices, it is impossible for engineers to explore each option. And since every application is different, it may not be appropriate to replicate solutions based on past implementations. However, by asking a few key questions and finding the right partner to take your project from ideation to silicon, any enterprise can develop a roadmap to successfully deploy edge AI in their devices.
Defining Use Cases And Feasibility
It is important to define use cases before exploring implementation options. The first question any business should ask is: What would the customer find truly useful? After pinpointing the functionality the customer wants, your team needs to set development and production cost goals along with the acceptable time to market.
Now comes the challenging part – making technology-related decisions. Is it possible to implement that functionality within the cost, time, power and space trade offs you’re dealing with? Working with an experienced partner/consultant or drawing on internal experience is critical at this stage. You won’t have perfect data on which to base your decision, so actual experience is essential in making these judgements.
Technology Choice
There are several choices a team can make to implement the customer’s desired functionality in the product. Depending on your available resources and development time, here are some of the choices a team might consider:
It can be difficult to evaluate the right choice without expertise from trained professionals in the edge AI chip space. Evaluation of each option can often take a long time and require extensive knowledge. For example, evaluating a fixed accelerator versus an FPGA implementation can require engineers with different skill sets. With so many vendors and solutions making conflicting claims, making basic decisions can be overwhelming for most enterprises.
One of the most important steps one can take is to find the right partner who can help evaluate the technology trade offs and take the company from the initial research and evaluation stage to the design and implementation of the solution. Also, don’t get hung up on finding the solution with the “optimum” power/performance. If you can identify a solution that will work and has adequate software and technical support, that is likely your best choice. Don’t get caught chasing specs.
Building The Solution
Once functionality and technology have been chosen, the next step is implementation. Often the focus is on the implementation of a neural network model, however businesses also have to deal with the implementation of logic (software/hardware) to handle the pipeline from sensor to final output, requiring unique algorithms at each step.
Questions that might come up include:
With so many vendors voicing conflicting claims, it is important for businesses looking to implement edge AI not to focus on finding the “best TOPS” or the “fastest” solution, as these are elusive goals. The best way to answer many of these questions of functionality, technology choice and implementation is to partner with a person or organization that has “been there, done that.” Someone with the experience to quickly evaluate potential use cases, technical solutions and vendor offerings to help you narrow your choices as quickly and accurately as possible. Focus on vendors that have the most complete solution, with both the engine to implement, but also models, algorithms, and even existing data to help you in your unique use case and create a solid proof of concept.
Implementing edge AI: Look before you leap
June 28, 2022 Douglas FairbairnAs the need for artificial intelligence grows more common and technology needs become more sophisticated, companies looking to adopt edge AI into their products often find it to be a difficult challenge. But what makes it so difficult, and what solutions exist to solve this problem?
Perhaps the single biggest issue that companies face in implementing edge AI is that most companies don’t have the resources in house to develop these sophisticated fast-changing technologies. Lack of trained personnel and little familiarity with design flow often leads to delayed timelines and excess expense to train team members. In addition, there are so many choices, it is impossible for engineers to explore each option. And since every application is different, it may not be appropriate to replicate solutions based on past implementations. However, by asking a few key questions and finding the right partner to take your project from ideation to silicon, any enterprise can develop a roadmap to successfully deploy edge AI in their devices.
Defining Use Cases And Feasibility
It is important to define use cases before exploring implementation options. The first question any business should ask is: What would the customer find truly useful? After pinpointing the functionality the customer wants, your team needs to set development and production cost goals along with the acceptable time to market.
Now comes the challenging part – making technology-related decisions. Is it possible to implement that functionality within the cost, time, power and space trade offs you’re dealing with? Working with an experienced partner/consultant or drawing on internal experience is critical at this stage. You won’t have perfect data on which to base your decision, so actual experience is essential in making these judgements.
Technology Choice
There are several choices a team can make to implement the customer’s desired functionality in the product. Depending on your available resources and development time, here are some of the choices a team might consider:
- Software only on the existing embedded processor – This may require very carefully coded models in order to achieve the desired performance. Functionality may be limited, but it is generally the lowest cost solution if it works. Because this is a software-only solution, upgrades or bug fixes are more easily addressed.
- Upgrade/replace the existing processor – This can be a great solution if you can make it work and preserve existing code base, and like the solution above, is software-only and can be easily fixed or upgraded. However, this can often start a project down a slippery path that requires extensive power and performance evaluation. Companies may be better off adding a neural network (NN) or similar accelerator.
- Add a fixed neural network accelerator – This is an optimum choice if there is a good match with the needs of the application, as evaluation and design may not be too difficult. It could very well provide excellent power/performance trade offs at a very reasonable cost.
- FPGA – This solution is flexible and upgradable, but typically comes with high cost and high power for the final product. Rarely is this a good choice for “edge” products.
- Dedicated SoC – Often this is the optimum choice for high volume, low cost and low power products where use cases are clearly defined.
It can be difficult to evaluate the right choice without expertise from trained professionals in the edge AI chip space. Evaluation of each option can often take a long time and require extensive knowledge. For example, evaluating a fixed accelerator versus an FPGA implementation can require engineers with different skill sets. With so many vendors and solutions making conflicting claims, making basic decisions can be overwhelming for most enterprises.
One of the most important steps one can take is to find the right partner who can help evaluate the technology trade offs and take the company from the initial research and evaluation stage to the design and implementation of the solution. Also, don’t get hung up on finding the solution with the “optimum” power/performance. If you can identify a solution that will work and has adequate software and technical support, that is likely your best choice. Don’t get caught chasing specs.
Building The Solution
Once functionality and technology have been chosen, the next step is implementation. Often the focus is on the implementation of a neural network model, however businesses also have to deal with the implementation of logic (software/hardware) to handle the pipeline from sensor to final output, requiring unique algorithms at each step.
Questions that might come up include:
- What kind of signal conditioning/filtering do I need before passing the data to a NN accelerator?
- Which NN model should I use? Is there an existing model for my technology selection? Which version of which model is best in my application?
- How do I train my model? Where do I get my data and what biases are built into that data? What volume of data do I need?
- What is the cost and availability of the processing power for training models? Do we train in the cloud or on local servers?
- What level of accuracy is adequate? Is it better to have false positives or false negatives?
- What post processing is required and can I handle that workload?
With so many vendors voicing conflicting claims, it is important for businesses looking to implement edge AI not to focus on finding the “best TOPS” or the “fastest” solution, as these are elusive goals. The best way to answer many of these questions of functionality, technology choice and implementation is to partner with a person or organization that has “been there, done that.” Someone with the experience to quickly evaluate potential use cases, technical solutions and vendor offerings to help you narrow your choices as quickly and accurately as possible. Focus on vendors that have the most complete solution, with both the engine to implement, but also models, algorithms, and even existing data to help you in your unique use case and create a solid proof of concept.
Douglas Fairbairn is a Silicon Valley veteran and currently director of business development for MegaChips, a $1 billion Japanese ASIC company expanding into the US. After graduating from Stanford with an MSEE, he spent 8 years at Xerox PARC. He then helped establish the ASIC business as cofounder of VLSI Technology and later of Redwood Design Automation, where he served as CEO until its acquisition by Cadence. He is now leveraging his ASIC and startup experience by helping establish MegaChips as a leading ASIC vendor in the US with special expertise in Edge AI technology. |
Interview 3 months ago with BRN CFO Ken Scarince.
Pretty bland interview but when asked about competition (around 2.30 min mark) he said we are disrupting an industry that does not exist yet. - AI at the Edge.
Around 2.50 mark still on competition he said Intel had just released their 2nd research chip and were 2 or 3 years behind.
Interesting.
Stable Genius
Regular
Hi @manny100
This interview is at least 2 years old!
Ken speaks well though, comes across as a direct straight talker.
JDelekto
Regular
Are we sure that Qualcomm's latest Snapdragon doesn't provide an SNN?
This old article from 2013 suggests that Qualcomm's Zeroth platform was using an SNN and I wonder if that technology has made it into the newer processors.
Tothemoon24
Top 20
This company is at the cutting edge in cancer detection.
Processing of data I can’t find atm , might be worth a closer look .
Sending much love to those who’s lives have been effected by this cruel monster .
Be great if the mighty Bchip played a part in kicking the f..ker in the arse
Back
Share this article:
Ibex’s AI Algorithm to Aid Pathologists with Accurate and Reproducible HER2 and HER2-low Assessment in Breast Cancer Diagnosis
Tel Aviv, Israel – January 24, 2023 – Ibex Medical Analytics (Ibex), the leader in AI-powered cancer diagnostics, today announced an agreement with AstraZeneca and Daiichi Sankyo, for the development, clinical validation and early adoption of an AI-powered product to aid pathologists with an accurate and reproducible assessment of HER2 immunohistochemistry (IHC) scoring in breast cancer patients.
Scoring of HER2 (human epidermal growth factor receptor 2) protein expression in breast cancer is used to identify patients who are likely to benefit from HER2-directed therapies. Currently, pathologists routinely score HER2 in tumor samples visually using a microscope, which can be challenging in cases of low HER2 expression as scoring is subjective and may lead to varied interpretations. Computational tools developed using Artificial Intelligence have the potential to support pathologists in accurate and objective scoring of HER2, which can help oncologists in selecting therapies that are approved for treating patients with HER2-positive or HER2-low breast cancer.
“Recognizing the vital role pathologists play in the diagnosis and treatment of cancer patients, we are thrilled to partner with AstraZeneca and Daiichi Sankyo to clinically validate our automated HER2 scoring product and offer it to laboratories around the world,” said Joseph Mossel, Co-founder and CEO of Ibex Medical Analytics. “As the most commonly diagnosed cancer in women, this collaboration will allow pathologists to utilize our technology to optimize breast cancer diagnosis and ultimately improve the identification of patients eligible for HER2-directed therapy.”
Ibex’s Galen™ Breast HER2 is an IHC scoring product that detects tumor areas and quantifies HER2 expression into four standard categories, 0, 1+, 2+ and 3+, based on the 2018 ASCO/CAP scoring guidelines1. As part of this collaboration, Ibex will work with AstraZeneca and Daiichi Sankyo to develop and clinically validate its HER2 IHC scoring product and generate evidence that further supports adoption of the technology.
A multi-site validation study on Galen Breast HER2 involved a cohort of 453 breast tumors of diverse subtypes. The study demonstrated that Galen’s AI algorithm provides an accurate and automated HER2 score for pathologists and was recently presented at the San Antonio Breast Cancer Symposium2.
Beyond this collaboration, Ibex supports pathologists with AI-based diagnostic solutions that help detect and grade different types of invasive and non-invasive breast cancer and other tumor types, and are used in everyday practice in laboratories, hospitals and health systems worldwide. Ibex’s Galen Breast solution demonstrated robust outcomes in detecting and grading multiple types of breast cancer and other clinically relevant features across clinical studies performed on numerous diagnostic workflows, one of which was recently published in Nature’s peer-reviewed npj Breast Cancer journal3,4.
In addition to HER2, Ibex is further expanding Galen Breast to include automated quantification of additional IHC-stained slides, such as ER, PR and Ki-67, intended to provide pathologists with a comprehensive set of tools for breast cancer diagnosis. With these expanded capabilities, Galen Breast may further enhance diagnostic efficiency and enable more accurate and objective scoring of breast biomarkers, improving treatment decisions and patient care.
NEWS PROVIDED BY
Ibex Medical Analytics
Feb 09, 2023, 08:00 ET
Pathology Diagnostics Platform Meets Safety, Quality and Performance Criteria Under EU's New In Vitro Diagnostic Medical Devices Regulation.
TEL AVIV, Israel, Feb. 9, 2023 /PRNewswire/ -- Ibex Medical Analytics (Ibex), the leader in AI-powered cancer diagnostics, today announced that Galen™ Prostate is now CE marked under the In Vitro Diagnostic Medical Devices Regulation (IVDR) for supporting pathologists in primary diagnosis of prostate biopsies. Galen Prostate is the first standalone AI-based cancer diagnostics product of its kind certified under the IVDR.
IVDR is the new regulatory standard set by the European Union, replacing the previous In Vitro Diagnostic Medical Device Directive (IVDD). The new regulation sets a new bar for product performance and clinical validation, as well as post marketing surveillance. Galen Prostate received its IVDR CE certificate following a rigorous review demonstrating the quality of the product and its meticulous development process, safety, and performance. During 2023, Ibex plans to migrate additional products, including its Galen Breast and Galen Gastric solutions, under the IVDR certificate.
"Ibex continues to maintain the highest possible standards for its products, bringing cutting-edge computational solutions to improve outcomes of cancer care," said Dr. Yael Liebes-Peer, Head of Regulatory Affairs and Quality Assurance at Ibex Medical Analytics. "Dedicated to our mission of providing every patient with an accurate, timely and personalized cancer diagnosis, we are proud to provide the market's first IVDR-certified product, elevating the quality of diagnosis for patients, pathologists and laboratories."
To help improve the quality of cancer diagnosis, increase productivity and optimize pathology workflows, Galen Prostate uses AI to analyze biopsies ahead of pathologists' review, providing them with diagnostic insights to guide their diagnosis. Galen Prostate's algorithms were trained on large datasets from multiple pathology institutes around the world, enriched with rare prostatic malignancies. Galen helps pathologists diagnose cancer, provides additional insights, including a Gleason score, tumor size and associated morphologies for each cancer slide, and offers decision support tools to help accelerate diagnostic turnaround and reduce subjectivity.
Ibex offers the most widely deployed AI technology in pathology, supporting pathologists worldwide with augmented diagnostic capabilities during diagnosis of breast, prostate, and gastric biopsies. Improving dignostic accuracy, reducing turnaround time, boosting productivity and improving user experience for pathologists, Galen has demonstrated excellent outcomes across multiple clinical studies performed in different pathology labs and diagnostic workflows1,2,3,4,5.
Processing of data I can’t find atm , might be worth a closer look .
Sending much love to those who’s lives have been effected by this cruel monster .
Be great if the mighty Bchip played a part in kicking the f..ker in the arse
- Ibex Medical Analytics Enters Collaboration with AstraZeneca and Daiichi Sankyo to Develop AI-based HER2 Scoring Product
Back
Share this article:
Ibex’s AI Algorithm to Aid Pathologists with Accurate and Reproducible HER2 and HER2-low Assessment in Breast Cancer Diagnosis
Tel Aviv, Israel – January 24, 2023 – Ibex Medical Analytics (Ibex), the leader in AI-powered cancer diagnostics, today announced an agreement with AstraZeneca and Daiichi Sankyo, for the development, clinical validation and early adoption of an AI-powered product to aid pathologists with an accurate and reproducible assessment of HER2 immunohistochemistry (IHC) scoring in breast cancer patients.
Scoring of HER2 (human epidermal growth factor receptor 2) protein expression in breast cancer is used to identify patients who are likely to benefit from HER2-directed therapies. Currently, pathologists routinely score HER2 in tumor samples visually using a microscope, which can be challenging in cases of low HER2 expression as scoring is subjective and may lead to varied interpretations. Computational tools developed using Artificial Intelligence have the potential to support pathologists in accurate and objective scoring of HER2, which can help oncologists in selecting therapies that are approved for treating patients with HER2-positive or HER2-low breast cancer.
“Recognizing the vital role pathologists play in the diagnosis and treatment of cancer patients, we are thrilled to partner with AstraZeneca and Daiichi Sankyo to clinically validate our automated HER2 scoring product and offer it to laboratories around the world,” said Joseph Mossel, Co-founder and CEO of Ibex Medical Analytics. “As the most commonly diagnosed cancer in women, this collaboration will allow pathologists to utilize our technology to optimize breast cancer diagnosis and ultimately improve the identification of patients eligible for HER2-directed therapy.”
Ibex’s Galen™ Breast HER2 is an IHC scoring product that detects tumor areas and quantifies HER2 expression into four standard categories, 0, 1+, 2+ and 3+, based on the 2018 ASCO/CAP scoring guidelines1. As part of this collaboration, Ibex will work with AstraZeneca and Daiichi Sankyo to develop and clinically validate its HER2 IHC scoring product and generate evidence that further supports adoption of the technology.
A multi-site validation study on Galen Breast HER2 involved a cohort of 453 breast tumors of diverse subtypes. The study demonstrated that Galen’s AI algorithm provides an accurate and automated HER2 score for pathologists and was recently presented at the San Antonio Breast Cancer Symposium2.
Beyond this collaboration, Ibex supports pathologists with AI-based diagnostic solutions that help detect and grade different types of invasive and non-invasive breast cancer and other tumor types, and are used in everyday practice in laboratories, hospitals and health systems worldwide. Ibex’s Galen Breast solution demonstrated robust outcomes in detecting and grading multiple types of breast cancer and other clinically relevant features across clinical studies performed on numerous diagnostic workflows, one of which was recently published in Nature’s peer-reviewed npj Breast Cancer journal3,4.
In addition to HER2, Ibex is further expanding Galen Breast to include automated quantification of additional IHC-stained slides, such as ER, PR and Ki-67, intended to provide pathologists with a comprehensive set of tools for breast cancer diagnosis. With these expanded capabilities, Galen Breast may further enhance diagnostic efficiency and enable more accurate and objective scoring of breast biomarkers, improving treatment decisions and patient care.
Ibex's Galen™ Prostate Becomes First Standalone AI-Powered Cancer Diagnostics Solution to Obtain CE Mark Under the IVDR
NEWS PROVIDED BY
Ibex Medical Analytics
Feb 09, 2023, 08:00 ET
Pathology Diagnostics Platform Meets Safety, Quality and Performance Criteria Under EU's New In Vitro Diagnostic Medical Devices Regulation.
TEL AVIV, Israel, Feb. 9, 2023 /PRNewswire/ -- Ibex Medical Analytics (Ibex), the leader in AI-powered cancer diagnostics, today announced that Galen™ Prostate is now CE marked under the In Vitro Diagnostic Medical Devices Regulation (IVDR) for supporting pathologists in primary diagnosis of prostate biopsies. Galen Prostate is the first standalone AI-based cancer diagnostics product of its kind certified under the IVDR.
IVDR is the new regulatory standard set by the European Union, replacing the previous In Vitro Diagnostic Medical Device Directive (IVDD). The new regulation sets a new bar for product performance and clinical validation, as well as post marketing surveillance. Galen Prostate received its IVDR CE certificate following a rigorous review demonstrating the quality of the product and its meticulous development process, safety, and performance. During 2023, Ibex plans to migrate additional products, including its Galen Breast and Galen Gastric solutions, under the IVDR certificate.
"Ibex continues to maintain the highest possible standards for its products, bringing cutting-edge computational solutions to improve outcomes of cancer care," said Dr. Yael Liebes-Peer, Head of Regulatory Affairs and Quality Assurance at Ibex Medical Analytics. "Dedicated to our mission of providing every patient with an accurate, timely and personalized cancer diagnosis, we are proud to provide the market's first IVDR-certified product, elevating the quality of diagnosis for patients, pathologists and laboratories."
To help improve the quality of cancer diagnosis, increase productivity and optimize pathology workflows, Galen Prostate uses AI to analyze biopsies ahead of pathologists' review, providing them with diagnostic insights to guide their diagnosis. Galen Prostate's algorithms were trained on large datasets from multiple pathology institutes around the world, enriched with rare prostatic malignancies. Galen helps pathologists diagnose cancer, provides additional insights, including a Gleason score, tumor size and associated morphologies for each cancer slide, and offers decision support tools to help accelerate diagnostic turnaround and reduce subjectivity.
Ibex offers the most widely deployed AI technology in pathology, supporting pathologists worldwide with augmented diagnostic capabilities during diagnosis of breast, prostate, and gastric biopsies. Improving dignostic accuracy, reducing turnaround time, boosting productivity and improving user experience for pathologists, Galen has demonstrated excellent outcomes across multiple clinical studies performed in different pathology labs and diagnostic workflows1,2,3,4,5.
Zedjack33
Regular
Tothemoon24
Top 20
So sorry to read your unfortunate news Zedjack, wishing you & your brother the best of luck .
Rise from the ashes
Regular
Best of luck to your brother my thoughts and prayers are with him. @Zedjack33
Hi JD,
A few days ago, @Stable Genius posted this diagram of Snapdragon 8.2 which has a pair of AI preprocessors:
In response, I posted this Qualcomm patent application which shows an arrangement with a pair of "split" accelerators:
US2020250545A1 SPLIT NETWORK ACCELERATION ARCHITECTURE
[0041] FIG. 3 is a block diagram illustrating a neural network acceleration architecture 300 , in accordance with aspects of the present disclosure. The neural network acceleration architecture 300 includes a first AI inference accelerator (e.g., first AIIA 330 ), a second AI inference accelerator (e.g., second AIIA 340 ), and a host processor 310 . In this configuration, the host processor 310 includes a host runtime block 312 to execute a host application 320 to operate a neural network. In this example, the neural network of the host application 320 exceeds a fixed amount of memory provided by a single AI inference accelerator (AIIA).
[0042] According to this aspect of the present disclosure, the neural network of the host application 320 is split across the first AIIA 330 and the second AIIA 340 .
Interestingly, the block diagram of the Snapdragon in the patent is virtually unchanged from the block diagram in the 2013 article.
Jumpchooks
Regular
Phobile Mones?
@Diogenese your assistance please with this Intel patent.
Are they trying to pole vault the moat and block Brainchip's future patents using external memory.
https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/11593623
Thanks in advance.
Are they trying to pole vault the moat and block Brainchip's future patents using external memory.
https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/11593623
Thanks in advance.
Frangipani
Regular
@Zedjack33, sorry to hear that your family is going through this… All the best for your brother!
Hopefully mRNA technology will - again - prove to be a game changer with personalised therapeutic cancer vaccines in the near future. While the COVID-19 pandemic abruptly thrust mRNA technology into the limelight on a global stage and most people had probably never heard of it prior to 2020, oncological research in this field has actually been going on for two decades and recent trials have shown promising results.
Hopefully mRNA technology will - again - prove to be a game changer with personalised therapeutic cancer vaccines in the near future. While the COVID-19 pandemic abruptly thrust mRNA technology into the limelight on a global stage and most people had probably never heard of it prior to 2020, oncological research in this field has actually been going on for two decades and recent trials have shown promising results.
Not sure if this has been shared but it’s great to see Global Foundries, GM, Renesas, Oculi and Brainchip mentioned in the same article.
I particularly liked this:
“GF is also championing the computing sector, namely with BrainChip's Akida neuromorphic chip built on 22 nm fully depleted silicon-on-insulator (FD-SOI)technology.”
I particularly liked this:
“GF is also championing the computing sector, namely with BrainChip's Akida neuromorphic chip built on 22 nm fully depleted silicon-on-insulator (FD-SOI)technology.”
Hi MrRomper,
That patent is no threat to Akida.
It has very inelegant architecture and uses processors during the inference operation. It also is synchronous.
All that interplay with external memories is an homage to John von Neumann.
The claims relate to a configuration which is alien to Akida:
1. A system for processing spiking neural network operations, the system comprising:
a plurality of neural processor clusters, each of the neural processor clusters to operate a plurality of neurons of a neural network, wherein each of the neural processor clusters comprises:
at least one neural processor to determine respective states of the plurality of neurons;
internal memory to maintain the respective states of the plurality of neurons;
an address generation unit to determine, based on a spike message received from a respective axon processor, a corresponding memory address for a particular postsynaptic neuron tracked in the internal memory; and
a control unit to:
retrieve neuron state data for the particular postsynaptic neuron from the internal memory, based on the corresponding memory address;
modify a potential of the particular postsynaptic neuron indicated in the neuron state data, based on the spike message received from the respective axon processor; and
write the neuron state data for the particular postsynaptic neuron back to the internal memory;
a plurality of axon processors, each of the axon processors to process synapse data of a plurality of synapses in the neural network, wherein the axon processors are coupled to respective banks of external memory, and wherein each of the axon processors comprises a control unit to:
retrieve synapse data of a subset of the plurality of synapses from a respective bank of the external memory;
evaluate the synapse data, based on a spike message received from a presynaptic neuron of a neural processor cluster; and
transmit, based on the evaluated synapse data, a weighted spike message to a postsynaptic neuron at a neural processor cluster.
This flow chart illustrates the use of the neuron processor (605) and the axon processor (610) as well as retrieving data from external memory (615) during operation.
Similar threads
