BRN Discussion Ongoing

Bravo

If ARM was an arm, BRN would be its biceps💪!
Nice find @Rocket577

It definitely shows that Nintendo are playing in our "sandbox". Still talking back propagation though.

I highlighted a paragraph I found interesting.


Systems and methods of neural network training​

Abstract​

A computer system is provided for training a neural network that converts images. Input images are applied to the neural network and a difference in image values is determined between predicted image data and target image data. A Fast Fourier Transform is taken of the difference. The neural network is trained on based the L1 Norm of resulting frequency data.


  • TECHNICAL OVERVIEW
  • [0002]
    The technology described herein relates to machine learning and training machine learned models or systems. More particularly, the technology described includes subject matter that relates to training neural networks to convert or upscale images by using, for example, Fourier Transforms (FTs).
  • INTRODUCTION
  • [0003]
    Machine learning can give computers the ability to “learn” a specific task without expressly programming the computer for that task. An example of machine learning systems includes deep learning neural networks. Such networks (and other forms of machine learning) can be used to, for example, help with automatically recognizing whether a cat is in a photograph. The learning takes place by using thousands or millions of photos to “train” the network (also called a model) to recognize when a cat is in a photograph. The training process can include, for example, determining weights for the model that achieve the indicated goal (e.g., identifying cats within a photo). The training process may include using a loss function in a way (e.g., via backpropagation) that seeks to train the model or neural network that will minimize the loss represented by the function. Different loss functions include L1 (Least Absolute Deviations) and L2 (Least Square Errors) loss functions.
  • [0004]
    It will be appreciated that new and improved techniques, systems, and processes are continually sought after in these areas of technology, such as technology that is used to train machine learned models or neural networks.
  • SUMMARY
  • [0005]
    In some examples, computer system for training a neural network that processes images is provided. In some examples, the system is used to train neural networks to upscale images from one resolution to another resolution. The system may include computer storage that stores image data for a plurality of images. The system may be configured to generate, from the plurality of images, input image data and then apply the input image data to a neural network to generate predicted output image data. A difference between the predicted output image data and target image data is calculated and that difference may then be transformed in frequency domain data. In some examples, the L1-loss is then used on the frequency domain data calculated from the difference, which is then used to train the neural network using backpropagation (e.g., stochastic gradient descent). Using the L1 loss may encourage sparsity of the frequency domain data, which may also be referred to, or part of, Compressed Sensing. In contrast, using the L2 loss on the same frequency domain data may generally not produce in good results due to, for example, Parseval's Theorem. In other words, the L2 loss of frequency transformed data, using a Fourier Transform, is the same as the L2 loss of the data—i.e., L2(FFT(data))=L2(data). Thus, frequency transforming the data when using an L2 loss may not produce different results.

TT


The patent refers to ASICs. And MegaChips is now delivering its full-service ASIC solution in the US and offering off-the-shelf access to industry-standard IP components, such as ours. So couldn't they just squish AKIDA 1500 into the ASIC so the existing patent would still be applicable?

PS: I'm not an engineer and I frequently have no idea what I'm talking about.



ASIC 2 pm.png

Screen Shot 2023-06-16 at 1.11.24 pm.png

ASIC pm.png
Screen Shot 2023-06-16 at 1.11.50 pm.png
 
  • Like
  • Love
  • Fire
Reactions: 44 users

Yoda

Regular
It's all slowly coming together and in the not so distant future I believe the days of little or no revenue will be a distant memory . . . :)
 
  • Like
  • Love
Reactions: 35 users

IloveLamp

Top 20

“Our client roadmap demonstrates how Intel is prioritizing innovation and technology leadership with products like Meteor Lake, focused on power efficiency and AI at scale. To better align with our product strategies, we are introducing a branding structure that will help PC buyers better differentiate the best of our latest technology and our mainstream offerings.”

–Caitlin Anderson, Intel vice president and general manager of Client Computing Group Sales
 
  • Like
  • Fire
Reactions: 6 users
D

Deleted member 118

Guest
  • Haha
Reactions: 2 users
Very Interesting
Screenshot_20230616-141605_Samsung Internet.jpg
 
  • Like
  • Fire
Reactions: 19 users

Diogenese

Top 20
  • Haha
Reactions: 18 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
My carrots are bigger than yours! 🥕


200w.gif
 
  • Haha
  • Like
  • Sad
Reactions: 12 users

IloveLamp

Top 20
The headline reminds me of the ant floating on a leaf down the river on his back with an erection shouting "Raise the drawbridge!"
wtf.gif
 
  • Haha
  • Like
Reactions: 8 users

IloveLamp

Top 20




Screenshot_20230616_175025_LinkedIn.jpg
 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 25 users

Pepsin

Regular
Interesting post on linkedin?! Liked by 2.500 people so far. For me it shows that neuromorphic vision is on the rise!
Is there a place for Akida in the arcitecture shown in the video?
1686901451997.png

1686901833802.png
 
  • Like
  • Fire
Reactions: 14 users

Easytiger

Regular
Brainchip advanced neuromorphic IP in megachips!

View attachment 38436

Provision of Edge AI Subsystem

We co-design subsystem incorporating characteristic Edge AI IP according to the customer’s system. (We can also support specified AI IP.)

  • Example 1: Intuitive UI subsystem (BrainChip IP)
  • Example 2: Image AI subsystem (Quadric IP)

Revenue forecast?
 
  • Like
  • Haha
  • Fire
Reactions: 5 users

Fox151

Regular
  • Haha
  • Like
  • Love
Reactions: 12 users

HopalongPetrovski

I'm Spartacus!
Revenue forecast?
I doubt we'll get anything like a revenue forecast from Sean until they have at least a few quarters of reliable and consistent data upon which to base it on and then project. But yes, especially if its Nintendo related and they can once again come up with something compelling and that catches the markets imagination then it could well be drool worthy. I think PVDM suggested at one point that Megachips could be huge for us.
 
  • Like
  • Love
  • Fire
Reactions: 35 users

IloveLamp

Top 20
  • Like
  • Fire
  • Thinking
Reactions: 12 users
D

Deleted member 118

Guest


Socionext to Showcase New Automotive Radar Sensor Technology at Sensors Converge 2023, Booth 1549​

Share this post
Share TweetShare

Advanced RF CMOS Sensors for In-Cabin Driver and Passenger Monitoring Systems Deliver Ground-breaking Functional and Safety Benefits​

Milpitas, Calif., Jun 8, 2023 --- Socionext, a global leader in high-precision sensor technology, will showcase its cutting-edge millimeter-wave integrated radar solutions for optimizing in-cabin automotive applications at the annual Sensors Converge Conference & Expo 2023 at the Santa Clara (Calif.) Convention Center June 21-22.
At Booth 1549, Socionext will debut its new AEC-Q100 automotive-grade qualified SC1260 Series intelligent, ultra-low-power, ultra-compact, all-in-one CMOS 60GHz radar sensors. The devices are designed to enable easy acquisition of three-dimensional (3D) positions, and feature temperature tolerances ranging from -40°C to 125°C.
One of the members of the new series, the SC1260AR3, has been nominated as a “2023 Best of Sensors Award” finalists. The SC1260AR3 comes with time-division multiplexing (TDM-MIMO) operation and multiple transmitting and receiving Antennas-in-Package (AiP) that can very accurately detect the position and movement of multiple passengers in a vehicle.
The new SC1260 Series offers a broad range of benefits for automotive applications, including:
60GHz radar sensors for high-precision occupant detection and infant wellness monitoring.
The device is explicitly designed for high-precision in-cabin sensing with TDM-MIMO, enabling a single sensor to detect multiple passengers seated in a row. The sensor has the capability to differentiate pets from humans and monitor vital signs such as heartbeat and respiration. Such technology is especially significant when infants and pets are left or forgotten in poorly ventilated or hot cars. Since radar can penetrate through solid non-metallic material, the technology can detect infants wrapped in blankets or hidden from view.
Anti-theft measures and monitoring of abnormal occurrences surrounding the vehicle, including pre- and post-collision video recording.
Radar has the capability to sense suspicious activities in and around the near field of a vehicle, then activate a dashcam to initiate recording. The ultra-low power sensors also can be useful in a vehicle that is parked and may be prone to hit-and-run incidents. While conventional camera systems record an intrusion or a collision only after they take place, new automotive dashcam systems using Socionext’s radar sensor can detect incoming vehicles and record an incident prior to the occurrence. The device is capable of 24/7 uninterrupted operation using a mere ~1mW of power.
Advances in touchless hand gesture control using the latest radar technology.
Socionext’s smart sensor offers 3D hand gesture recognition to operate in-vehicle displays and infotainment systems.
As the demand for high-precision, multi-functional automotive systems continues to grow, radar sensing technology products offered by Socionext provide a broad range of features and benefits for improving vehicle occupant safety and enhancing a car owner’s overall driving experience.
 
  • Like
  • Fire
  • Love
Reactions: 25 users
D

Deleted member 118

Guest
 
  • Like
  • Fire
  • Love
Reactions: 10 users

Slymeat

Move on, nothing to see.
Discussion - Automotive Supply Chain Challanges

Panel: “So Mr Hehir, what have your recent automotive supply chain challenges been?”

Sean Hehir: “Yeah, we’re finding it a challenge to find an automotive manufacturer to supply our product to… so yeah…”

JK, kinda 😂😳
I am embarrassed by that reply!
 
  • Like
  • Thinking
Reactions: 7 users

robsmark

Regular
I am embarrassed by that reply!
Don’t worry, I’m not offended.
 
  • Haha
  • Like
Reactions: 3 users
Top Bottom