BRN Discussion Ongoing

[tjcov87]

 

[Damo4]

Wow. Not sure what else to say, but the identification of the shark species is fantastic.
Don't care who's tech this is, I'm impressed.

It saw the bull shark before I did.

 

[Damo4]

Wow. Not sure what else to say, but the identification of the shark species is fantastic.
Don't care who's tech this is, I'm impressed.

It saw the bull shark before I did.

Was MobileNet V1 one of the things listed somewhere very recently in regards to Akida? On either the benchmarking paper or something released recently regarding platforms?


"Over the last five years remotely piloted drones have become the tool of choice to spot potentially dangerous sharks in New South Wales, Australia. They have proven to be a more effective, accessible and cheaper solution compared to crewed aircraft. However, the ability to reliably detect and identify marine fauna is closely tied to pilot skill, experience and level of fatigue. Modern computer vision technology offers the possibility of improving detection reliability and even automating the surveillance process in the future. In this work we investigate the ability of commodity deep learning algorithms to detect marine objects in video footage from drones, with a focus on distinguishing between shark species. This study was enabled by the large archive of video footage gathered during the NSW Department of Primary Industries Drone Trials since 2016. We used this data to train two neural networks, based on the ResNet-50 and MobileNet V1 architectures, to detect and identify ten classes of marine object in 1080p resolution video footage. Both networks are capable of reliably detecting dangerous sharks: 80% accuracy for RetinaNet-50 and 78% for MobileNet V1 when tested on a challenging external dataset, which compares well to human observers. The object detection models correctly detect and localise most objects, produce few false-positive detections and can successfully distinguish between species of marine fauna in good conditions. We find that shallower network architectures, like MobileNet V1, tend to perform slightly worse on smaller objects, so care is needed when selecting a network to match deployment needs. We show that inherent biases in the training set have the largest effect on reliability. Some of these biases can be mitigated by pre-processing the data prior to training, however, this requires a large store of high resolution images that supports augmentation. A key finding is that models need to be carefully tuned for new locations and water conditions. Finally, we built an Android mobile application to run inference on real-time streaming video and demonstrated a working prototype during fields trials run in partnership with Surf Life Saving NSW."

 

[Diogenese]

Was MobileNet V1 one of the things listed somewhere very recently in regards to Akida? On either the benchmarking paper or something released recently regarding platforms?


"Over the last five years remotely piloted drones have become the tool of choice to spot potentially dangerous sharks in New South Wales, Australia. They have proven to be a more effective, accessible and cheaper solution compared to crewed aircraft. However, the ability to reliably detect and identify marine fauna is closely tied to pilot skill, experience and level of fatigue. Modern computer vision technology offers the possibility of improving detection reliability and even automating the surveillance process in the future. In this work we investigate the ability of commodity deep learning algorithms to detect marine objects in video footage from drones, with a focus on distinguishing between shark species. This study was enabled by the large archive of video footage gathered during the NSW Department of Primary Industries Drone Trials since 2016. We used this data to train two neural networks, based on the ResNet-50 and MobileNet V1 architectures, to detect and identify ten classes of marine object in 1080p resolution video footage. Both networks are capable of reliably detecting dangerous sharks: 80% accuracy for RetinaNet-50 and 78% for MobileNet V1 when tested on a challenging external dataset, which compares well to human observers. The object detection models correctly detect and localise most objects, produce few false-positive detections and can successfully distinguish between species of marine fauna in good conditions. We find that shallower network architectures, like MobileNet V1, tend to perform slightly worse on smaller objects, so care is needed when selecting a network to match deployment needs. We show that inherent biases in the training set have the largest effect on reliability. Some of these biases can be mitigated by pre-processing the data prior to training, however, this requires a large store of high resolution images that supports augmentation. A key finding is that models need to be carefully tuned for new locations and water conditions. Finally, we built an Android mobile application to run inference on real-time streaming video and demonstrated a working prototype during fields trials run in partnership with Surf Life Saving NSW."

Hi Damo,

MobileNet is an open source model library database used to test and train NNs.

One of the stats re Akida you may have seen is the time it takes to classify the images in the library.

There are various versions adapted for different subject matter.

 

[Damo4]

Hi Damo,

MobileNet is an open source model library database used to test and train NNs.

One of the stats re Akida you may have seen is the time it takes to classify the images in the library.

There are various versions adapted for different subject matter.

Thank you @Diogenese I knew I had seen it somewhere!

 

[Diogenese]

Tough times for tech:

https://www.msn.com/en-au/news/tech...pc=U531&cvid=4d53fcc86fe24e50857948dfedf6763c

Microsoft set to lay off thousands of employees tomorrow​

Story by Tom Warren • 8h ago

Microsoft is preparing to announce job cuts tomorrow. Sky News reports that thousands of roles will be cut, with the software giant said to be looking at cutting around 5 percent of its workforce. With more than 220,000 employees at Microsoft, that could mean more than 10,000 layoffs.
...
The cuts also come just weeks after Microsoft CEO Satya Nadella warned of two years of challenges ahead for the tech industry. In an interview with CNBC, Nadella admitted Microsoft wasn’t “immune to the global changes” and spoke of the need for tech companies to be efficient.

“The next two years are probably going to be the most challenging,” said Nadella. “We did have a lot of acceleration during the pandemic, and there’s some amount of normalization of that demand. And on top of it, there is a real recession in some parts of the world.”

 

[Rskiff]

Tough times for tech:

https://www.msn.com/en-au/news/tech...pc=U531&cvid=4d53fcc86fe24e50857948dfedf6763c

Microsoft set to lay off thousands of employees tomorrow​

Story by Tom Warren • 8h ago

Microsoft is preparing to announce job cuts tomorrow. Sky News reports that thousands of roles will be cut, with the software giant said to be looking at cutting around 5 percent of its workforce. With more than 220,000 employees at Microsoft, that could mean more than 10,000 layoffs.
...
The cuts also come just weeks after Microsoft CEO Satya Nadella warned of two years of challenges ahead for the tech industry. In an interview with CNBC, Nadella admitted Microsoft wasn’t “immune to the global changes” and spoke of the need for tech companies to be efficient.

“The next two years are probably going to be the most challenging,” said Nadella. “We did have a lot of acceleration during the pandemic, and there’s some amount of normalization of that demand. And on top of it, there is a real recession in some parts of the world.”

And BRN is still on the hunt hiring.

 

[wilzy123]

And BRN is still on the hunt hiring.

Times are still good for some!

 

[thelittleshort]

Steven Peters from the Institute of Automotive Engineering (FZD) confirmed that they are doing some "work"/studies into Neuromorphic. Hopefully with Akida.

View attachment 27386

It’s better than that @alwaysgreen. Steve Peters was the Lead AI Researcher for Mercedes Benz on the Vision EQXX!

 

[Diogenese]

Tough times for tech:

https://www.msn.com/en-au/news/tech...pc=U531&cvid=4d53fcc86fe24e50857948dfedf6763c

Microsoft set to lay off thousands of employees tomorrow​

Story by Tom Warren • 8h ago

Microsoft is preparing to announce job cuts tomorrow. Sky News reports that thousands of roles will be cut, with the software giant said to be looking at cutting around 5 percent of its workforce. With more than 220,000 employees at Microsoft, that could mean more than 10,000 layoffs.
...
The cuts also come just weeks after Microsoft CEO Satya Nadella warned of two years of challenges ahead for the tech industry. In an interview with CNBC, Nadella admitted Microsoft wasn’t “immune to the global changes” and spoke of the need for tech companies to be efficient.

“The next two years are probably going to be the most challenging,” said Nadella. “We did have a lot of acceleration during the pandemic, and there’s some amount of normalization of that demand. And on top of it, there is a real recession in some parts of the world.”

"acceleration during the pandemic" - Looks like they are "right-sizing" after the boost that lockdowns and working from home gave to the interweb business.

 

[Fullmoonfever]

Talking of Cadence, they're also a sponsor of the upcoming DesignCon happening from Jan 31 to Feb 2:

DesignCon | January 28–30, 2025

DesignCon is the must attend event for all aspects of electronic design.

www.designcon.com

Also just on the Cadence comments.

I noticed this the other day.

Not so much directly about Akida but interrelationships with partners.

Socionext who were discussing their Automotive Graphics Display Controller recently appear to have utilised Cadences Stratus HSL in the design work.

Like to believe that in these interrelationships and tech cross overs that some cross pollination of ideas etc sometimes get discussed at an informal level.

In this Expert Insights Video, Socionext’s Tim Papenfuss discusses how and why they used SystemC and Stratus high-level synthesis (HLS) to design their SC1701 automotive graphics display controller. A explanation of Clock Domain Crossing in Stratus HLS is outlined.

https://www.cadence.com/ja_JP/home/tools/digital-design-and-signoff/synthesis/stratus-high-level-synthesis.html

Intel and Qualcomm also utilise Stratus and have some vids.

 

[SERA2g]

23 is my favourite number. If we were bought out for $23, I would definitely be building these (gaudy) front gates on my house!

View attachment 27382

They are Michael Jordans front gates btw. Money definitely doesn't buy class haha

I definitely consider $23 per share takeover for at least 5 minutes. Would toss a coin in the end.

$23 would get me a lambo but it wouldn't get me to the moon.

 

[kayaker]

While looking what the Chinese Tianjic chip is all about,which can work effectively with both ANN and SNN,I came across attached pdf which discusses a whole range of neural network chips including Akida.
It is quite educational and also pointing out where most of them fall short.

My apologies if attached article has been posted before.

 

[Rise from the ashes]

Ok now it's time for me to fully watch the latest podcast
Only managed to watch up to 1min 20.
Also avoided any comments about it here. So it's going to feel like a world premiere for me
And as usual I like to throw a bit of comedy in with my posts
Was chatting with my dog and out of nowhere he said he'd like to contribute to his upkeep. I was first shocked that he could speak but was curious how he could help support his lavish lifestyle.
He said he was applying for a job as Roofer.

 

[SERA2g]

The Valeo expert commeneted on the post stating "I can confirm that it is a long way to go but very promising". That doesn't fill me with any confidence that we will be in any Valeo products anytime soon.

Consider the context before you decide on how the statement impacts on your confidence in Valeo.

As per Markus' post on LinkedIn:

"The idea is not new, but trying to “put a brain on a chip” is a mammoth task. To put it into figures: the human brain has 86-100 billion neurons operating on around 20 watts. Current neural chips from leading developers such as BrainChip and Intel Corporation contain around 1 million neurons and consume roughly 1 watt of power."

Brainchip's akida is what many consider 'science fiction', yet it's 1 million neurons operating at 1 watt of power is only 0.001% of the human brain's 100 billion neurons.

There is still a long way to go in order to "put a human brain on a chip".

The good news though, Brainchip is one of the 'leading developers'.

Call me a glass-half-full kind of guy.

Edit: I'd like to expand on the glass-half-full comment.

Another way to think of Markus' post is as follows:

The Chief Technology Officer of Mercedes, a company at the cutting edge of automotive technology development, has just told you in plain written english that our small little 'meme-stock' start-up from modest beginnings here in Perth is rubbing shoulders with Intel, one of the worlds biggest technology company's, when it comes to the development and leadership in the field of neural chips.

If that doesn't fill you with confidence, I don't know what will.

 

[HopalongPetrovski]

Consider the context before you decide on how the statement impacts on your confidence in Valeo.

As per Markus' post on LinkedIn:

"The idea is not new, but trying to “put a brain on a chip” is a mammoth task. To put it into figures: the human brain has 86-100 billion neurons operating on around 20 watts. Current neural chips from leading developers such as BrainChip and Intel Corporation contain around 1 million neurons and consume roughly 1 watt of power."

Brainchip's akida is what many consider 'science fiction', yet it's 1 million neurons operating at 1 watt of power is only 0.001% of the human brain's 100 billion neurons.

There is still a long way to go in order to "put a human brain on a chip".

The good news though, Brainchip is one of the 'leading developers'.

Call me a glass-half-full kind of guy.

Journey of a Thousand miles begins with..........................AKIDA.

 

[chapman89]

Consider the context before you decide on how the statement impacts on your confidence in Valeo.

As per Markus' post on LinkedIn:

"The idea is not new, but trying to “put a brain on a chip” is a mammoth task. To put it into figures: the human brain has 86-100 billion neurons operating on around 20 watts. Current neural chips from leading developers such as BrainChip and Intel Corporation contain around 1 million neurons and consume roughly 1 watt of power."

Brainchip's akida is what many consider 'science fiction', yet it's 1 million neurons operating at 1 watt of power is only 0.001% of the human brain's 100 billion neurons.

There is still a long way to go in order to "put a human brain on a chip".

The good news though, Brainchip is one of the 'leading developers'.

Call me a glass-half-full kind of guy.

Edit: I'd like to expand on the glass-half-full comment.

Another way to think of Markus' post is as follows:

The Chief Technology Officer of Mercedes, a company at the cutting edge of automotive technology development, has just told you in plain written english that our small little 'meme-stock' start-up from modest beginnings here in Perth is rubbing shoulders with Intel, one of the worlds biggest technology company's, when it comes to the development and leadership in the field of neural chips.

If that doesn't fill you with confidence, I don't know what will.

But, but, but, I demand revenue now, none of that means anything

 

[equanimous]

But, but, but, I demand revenue now, none of that means anything

 

[cosors]

Good morning everyone.

Can we stop sharing Motley Fool articles here?

BRN holders share it everywhere and then we complain that picklebro writes non stop about Brainchip.

They get maximum exposure with Brainchip when holders write and discuss their articles constantly. They have every reason to write it because we create the buzz, make a platform and share it far and wide.

Stop posting their bs articles here and then the buzz goes away and they'll find another company.

This is my "soft" opinion only, you can continue to share it if you wish, thought I'd make the suggestion .

I have tried and asked here not to post it for the very reason of not giving them this platform for the first time. That did not work. But never give up hope. Hopefully it will work next for the first time!

 

[Rise from the ashes]

In the future I'll be posting Dumblilbro's articles about BRN across a wide range of financial sites. Obviously when that occurs BRN will be a raging financial successful company and Ficklemydicklebro's future articles will be printed and used as toilet paper for the future MF readership .