BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Fullmoonfever
Top 20
Good to see some game developers starting recognise the future re neuromorphic.
Blog excerpt below. Not us specifically but the tech which we know we are at the forefront.
www.filamentgames.com
In our last blog on AI, we covered different types of AI models. But how do these digital brains actually learn and train themselves? In this blog, we’re going to take a closer look into the processes behind AI learning and training, as well as cover some future considerations and caveats when it comes to using AI.
This also means that the economic value, the trade secret or the recipe of AI, if you will, are the weights and the training datasets. The network of connections are not nearly as valuable from a business standpoint. This is why Meta and other companies have started sharing untrained models publicly. That doesn’t mean there aren’t plenty of open source weights as well, but they are usually inferior quality to what big companies with lots of investment in training time have. This is also why publications and authors are upset that their articles were scraped into a training dataset without compensation.
When a trained model generates a response from a stimulus, it’s called inference, akin to human thought processes. While inference is generally swift, Generative Models face computational hurdles due to their expansive output layers, driving the race for AI-capable chips.
There are two kinds of AI-helpful chips – the kind everyone has been using to accelerate training and inference have been GPUs – which happen to be the same exact chips in your graphics card in your PC, PS5, or XBox. If you think that’s strange, you are not alone. While GPUs coincidentally accelerate certain AI tasks, the future is in Neuromorphic Processors. These implement neurons in silicon. This means pre-trained AI models will be able to run inference entirely in hardware, but it’s likely training will still need to rely on classical techniques for longer, so don’t dump your NVIDIA stock just yet!
Blog excerpt below. Not us specifically but the tech which we know we are at the forefront.
How AI Models Learn and Train Themselves | Filament Games
Curious about how AI models actually learn and train themselves? Here's an in-depth look at the fascinating processes behind AI learning and training!
www.filamentgames.com
how ai models learn and train themselves
March 27, 2024 // Alex StoneIn our last blog on AI, we covered different types of AI models. But how do these digital brains actually learn and train themselves? In this blog, we’re going to take a closer look into the processes behind AI learning and training, as well as cover some future considerations and caveats when it comes to using AI.
Neural Networks and Pre-Trained Models
All modern AI models rely on Neural Networks, mirroring the structure of the human brain. These networks consist of simulated neurons connected in various configurations, each with assigned “weights” dictating their influence. These weights, crucial for learning, are determined through rigorous training against labeled datasets. Models like ChatGPT and DALL-E are pre-trained, with companies investing heavily in refining these weights and datasets, making them the linchpin of AI value.This also means that the economic value, the trade secret or the recipe of AI, if you will, are the weights and the training datasets. The network of connections are not nearly as valuable from a business standpoint. This is why Meta and other companies have started sharing untrained models publicly. That doesn’t mean there aren’t plenty of open source weights as well, but they are usually inferior quality to what big companies with lots of investment in training time have. This is also why publications and authors are upset that their articles were scraped into a training dataset without compensation.
When a trained model generates a response from a stimulus, it’s called inference, akin to human thought processes. While inference is generally swift, Generative Models face computational hurdles due to their expansive output layers, driving the race for AI-capable chips.
There are two kinds of AI-helpful chips – the kind everyone has been using to accelerate training and inference have been GPUs – which happen to be the same exact chips in your graphics card in your PC, PS5, or XBox. If you think that’s strange, you are not alone. While GPUs coincidentally accelerate certain AI tasks, the future is in Neuromorphic Processors. These implement neurons in silicon. This means pre-trained AI models will be able to run inference entirely in hardware, but it’s likely training will still need to rely on classical techniques for longer, so don’t dump your NVIDIA stock just yet!
about filament games
we create playful experiences that improve people’s lives
Founded in 2005, Filament Games is a full-service digital studio that specializes in learning game development on a for-hire basis. We’ve completed over 400 projects since our founding and have worked with some of the biggest names in education – folks like Amazon, Scholastic, Smithsonian, Oculus, National Geographic, PBS, Television Ontario (TVO), McGraw-Hill, and even the US Department of Education.
Lovinglife
Member
CHIPS
Regular
It is so strange, that my posts are sometimes in German though I post them in English. It is no secret that I am German, but who the heck translates my post and why? #Akida was that you???
Last edited:
CHIPS
Regular
Besides from the German, what is the BN forum??
Krustor
Regular
Börsenews Forum - also known as the German kindergarden Forum - so never mind
CHIPS
Regular
Thanks. Nope, that's not my place to play!
cosors
👀
I hope it is a GNSS device instead of just GPS and uses the EU's Galileo - now called HAS - High Accuracy Service. Not that it is 10 or 20m off.)
Sign Up | LinkedIn
500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.www.linkedin.com
Volkswagen's RooBadge AI for safer roads in Australia
Volkswagen has developed a new technology called RooBadge to reduce kangaroo collisions in Australia.
This innovative device uses artificial intelligence to deter kangaroos from approaching vehicles.
Here's how it works:
RooBadge replaces the existing front badge on Volkswagen trucks.
It uses GPS and machine learning to identify the kangaroo species in the area.
Based on this information, it emits targeted sounds, including bird calls, predator sounds, and kangaroo foot thumps, which are natural warning signs for them.
The sound is focused ahead of the car, maximizing its effectiveness.
Benefits:
Reduces the risk of kangaroo collisions, protecting both drivers and animals.
Provides valuable data through driver and wildlife agency reports, allowing for continuous improvement.
A universal version is in development, potentially deterring other animals like deer and being adaptable for global use.
RooBadge is a prime example of how AI can be used to create positive change and save lives.
What are your thoughts about this innovative?
From 1 January 2023, three frequency bands
Embracing the Cosmos: Akida's Role in Advancing Space Tech for Daily Life
AI and robotics are enabling safer and more efficient space exploration.
www.wevolver.com
LinkedIn Login, Sign in | LinkedIn
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.
Socionext US on LinkedIn: Advanced SoC Research for Automotive 28/12/2023
We're thrilled to announce our participation in an innovative automotive chiplet R&D project spearheaded by The New Energy and Industrial Technology…
Hey FJ, I don't think they were ever "kicked to the kerb" when the Company changed, or affirmed a strategy of selling IP..
If there were buyers for the volume of chips, that we have, I think they would have just sold them..
Pretty sure it was stated at the time, of being the start of a mass production run, so there should have been possibly a few 10s of thousands produced (although we don't know and will never know the yield percentage, with Anil on the job, I suspect it would be pretty high).
With the speed of change in technology, I think it was a wise move, to follow the IP route, but even better now, with a multi-pronged strategy.
Some keep banging on (yes I'm talking about you Iseki) that we need to produce an AKD2000 reference chip, but there is no need as the tech can be played with in simulation.
Before the AKD1000 ES and reference chips, anybody looking at the simulations, would have thought "Yeah looks incredible, but can you really do that on a chip?"..
We don't have that problem anymore, as the core technology has been proven, in 2 processes (AKD1500).
Nice Dingo....including being able to be produced out of 3 different foundries, namely TSMC, Global and Intel, in different formats, NOT
forgetting more foundry/s are to be named, as per Sean Hehir.
Regards...Tech
Flawless Defect Detection with Edge Impulse and BrainChip
Machine learning algorithms and edge computing platforms can perform automated defect detection in a simple and cost-effective manner, as this project demonstrates.
hubs.ly
toasty
Regular
Isn't this what we were supposed to be doing with Safran/Airbus some years ago??View attachment 60177 View attachment 60178Flawless Defect Detection with Edge Impulse and BrainChip
Machine learning algorithms and edge computing platforms can perform automated defect detection in a simple and cost-effective manner, as this project demonstrates.
Was just a joke!
It is so strange, that my posts are sometimes in German though I post them in English. It is no secret that I am German, but who the heck translates my post and why? #Akida was that you???
sometimes the brain is playing games! One time i started to write something here. I thought I wrote in English and realised it was German
HopalongPetrovski
I'm Spartacus!
But they don’t lie
‘No, I’m not a robot’: ChatGPT successor tricks worker into thinking it is human
GPT-4 convinced TaskRabbit worker to solve Captcha
ChatGPT successor tricks worker into thinking it’s human
GPT-4 convinced TaskRabbit worker to solve Captcha
View attachment 60174 View attachment 60175 View attachment 60176
Socionext US on LinkedIn: Advanced SoC Research for Automotive 28/12/2023
We're thrilled to announce our participation in an innovative automotive chiplet R&D project spearheaded by The New Energy and Industrial Technology…www.linkedin.com
Morning IloveLamp
BrainChip and Socionext expect Akida silicon in Q3
BrainChip and Socionext expect engineering samples of their neural net processor Akida in Q3. They are being made on a TSMC MPW run. BrainChip has
www.electronicsweekly.com
This from Apr 2020, so could that mean that Socionext has something to offer with Akida already built in?
BrainChip and Socionext expect Akida silicon in Q3
BrainChip and Socionext expect engineering samples of their neural net processor Akida in Q3. They are being made on a TSMC MPW run.
BrainChip and Socionext expect Akida silicon in Q3
BrainChip and Socionext expect Akida silicon in Q3
brainchip.com
Morning everyone, have a great day ❤
ndefries
Regular
this is all getting out of hand. I thought I was investing in a meme stock. I am very concerned this could become a lot more. I am not going to renew my MF subscription!!!
Similar threads
