BRN Discussion Ongoing

[Frangipani]

Uni Luxembourg’s SIGCOM (Signal Processing and Communications) research group, which is part of SnT (Interdisciplinary Centre for Security, Reliability and Trust), has a new lab, which their Akida Shuttle PC now calls home: the TelecomAI-Lab, led by Flor Ortiz.

“Unlike traditional communication laboratories, TelecomAI-Lab focuses on AI-native solutions for signal processing and network optimisation.”

There is also an interesting section on “Services offered”.

TelecomAI - Lab | Symeon Chatzinotas

New HW Lab at SIGCOM, aggregating our capabilities on AI accelerators and Neuromorphic processors.

www.linkedin.com

View attachment 77272

TelecomAI - Lab

The TelecomAI at the University of Luxembourg is used for research and testing of signal processing techniques for satellite and terrestrial communication systems.

www.uni.lu

View attachment 77271
View attachment 77273
View attachment 77274

The Akida Shuttle PC from the TelecomAI-Lab at Uni Luxembourg’s Interdisciplinary Centre for Security, Reliability and Trust (SnT) was recently presented to the interested public at a booth showcasing the SIGCOM Research Group’s neuromorphic research (utilising both Loihi and Akida) on the occasion of SnT Partnership Day:

#pday2025 | Eva Lagunas

🛰️ On May 7th we celebrated the SnT Partnership Day #pday2025, where with the help of Suleima B., and Flor Ortiz, we organized a fantastic booth to showcase the SIGCOM Research Group activities and lab-facilities related to neuromorphic computing. More info: https://lnkd.in/ef_nhqyz SnT...

www.linkedin.com

Personal Website of Dr. Eva Lagunas - TelecomAI-Lab

Description: Led by Dr. Flor Ortiz and Suleima Briseño, the TelecomAI-Lab focuses on the research and development of embedded Artificial Intelligence solutions designed to optimize and revolutionize wireless communications, incorporating neuromorphic computing and satellite applications. Objectives

sites.google.com

Partnership Day 2025: Powering AI innovation

Prime Minister Luc Frieden visiting SnT's partnership Day

www.uni.lu

“AI is everywhere in our daily lives, but also in scientific research. The SnT Partnership Day 2025 showcased concrete projects.

Artificial intelligence is no longer a distant concept. It’s reshaping how we work, communicate, and innovate. On 7 May 2025, SnT’s annual Partnership Day brought together 700 participants from research, industry, and government representatives at the European Convention Centre Luxembourg (ECCL) to explore how AI is moving from theory to practice.

SnT is the University of Luxembourg’s tech research hub. The event highlighted how research partnerships are delivering real-world impact (…).”

Partnership Day 2025

Explore this photo album by SnT University of Luxembourg on Flickr!

www.flickr.com

 

[HopalongPetrovski]

It strikes me as strange that Nanovue (NVU) can today make a price sensitive announcement on the ASX that they are in effect doing a science experiment on their edge device to prove it may be useful and efficient in drones.
Seems they are cutting our lunch somewhat.
Good for them but just seems to illustrate a double standard in what is acceptable conduct in regards ASX announcements.

 

[jrp173]

I simply posted what the ann was, but whatever dk

Unfortunately some people just refuse to see what is happening.

How any can be okay with what's going on is crazy. Our Chairman told bold faced lies at the AGM (not just about the domicile) and the board sat and said nothing.

If this is not has not caused huge warning bells, then people are not paying attention.

This behaviour speaks volumes about our management.

 

[7für7]

Morning Chippers,

Conversion = AU$0.2654376 per share.

Hopefully mooving on up.

Regards,
Esq.

That’s what I told you! I don’t trust aftermarket movements

 

[overpup]

I find it amusing that the Loihi2 board has a "fragile, do not touch" sign...
Pretty much says it all really

 

[Guzzi62]

Unfortunately some people just refuse to see what is happening.

How any can be okay with what's gong on is crazy. Our Chairman told bold faced lies at the AGM (not just about the domicile) and the board sat and said nothing.

If this is not has not caused huge warning bells, then people are not paying attention.

This behaviour speaks volumes about our management.

And you point that out to us every day.

It is what it is, sell and move on if you can't take it, is that so hard?

So the domicile is shelled for now, as in 2017, okay noted, won't lose a minutes sleep over it.

I am buying more shares when I can.

You go back on ignore, a waste of time reading your negative posts, like the HC, a lot on ignore over there, you should go over there, you fit right in, complaining day after day, pathetic!

 

[jrp173]

And you point that out to us every day.

It is what it is, sell and move on if you can't take it, is that so hard?

So the domicile is shelled for now, as in 2017, okay noted, won't lose a minutes sleep over it.

I am buying more shares when I can.

You go back on ignore, a waste of time reading your negative posts, like the HC, a lot on ignore over there, you should go over there, you fit right in, complaining day after day, pathetic!

Guzzi62 I thought you were putting me on ignore.. can't help yourself eh....you are too funny....

 

[jrp173]

Calling me blind, huh?

There are different scenarios Mr negative.

You go on ignore, I am not wasting anymore time on you.

Bye bye.

You couldn't even ignore my posts for one week...

 

[Mccabe84]

Dropped all the way down now back up

 

[Bravo]

Where's the defibrillator? We appear to be flatlining...

 

[Pom down under]

Where's the defibrillator? We appear to be flatlining...


View attachment 84778

Can it be lDA offloading more of the shares they need to sell? As reading here over the last week ain’t BRN meant to be getting a much higher price than current SP? So if it’s not them then I guess we will see a much higher short position be taken

 

[Stable Genius]

Where's the defibrillator? We appear to be flatlining...


View attachment 84778

It’s ASX wide due to a drop in credit rating.

Live: Australian stocks slip ahead of RBA rates decision tomorrow — ABC News

The Australian share market has fallen at the open on Monday, as the Reserve Bank's two-day meeting gets underway today ahead of its interest rates decision tomorrow.

stocks.apple.com

As always unless you have an urgency to sell this price action is just white noise until we get contracts and revenue.

Sean’s remuneration is tied to sales so I’m sure he’s doing his best.

With FG, BH, AFRL, Chelpis Mirle (CM), OHB Hellas, QV, Andes and RISC V etc all on the job I’m hopeful we’ll get some deals over the line eventually.

 

[7für7]

I don’t get it… there are Canadian trash companies that have nothing to show for, have changed their name a hundred times and rebranded themselves, yet their stock price is higher.

 

[Pom down under]

It’s ASX wide due to a drop in credit rating.


With FG, BH, AFRL, Chelpis Mirle (CM), OHB Hellas, QV, Andes and RISC V etc all on the job I’m hopeful we’ll get some deals over the line eventually.

 

[TECH]

The Akida Shuttle PC from the TelecomAI-Lab at Uni Luxembourg’s Interdisciplinary Centre for Security, Reliability and Trust (SnT) was recently presented to the interested public at a booth showcasing the SIGCOM Research Group’s neuromorphic research (utilising both Loihi and Akida) on the occasion of SnT Partnership Day:

#pday2025 | Eva Lagunas

🛰️ On May 7th we celebrated the SnT Partnership Day #pday2025, where with the help of Suleima B., and Flor Ortiz, we organized a fantastic booth to showcase the SIGCOM Research Group activities and lab-facilities related to neuromorphic computing. More info: https://lnkd.in/ef_nhqyz SnT...

www.linkedin.com

View attachment 84765

Personal Website of Dr. Eva Lagunas - TelecomAI-Lab

Description: Led by Dr. Flor Ortiz and Suleima Briseño, the TelecomAI-Lab focuses on the research and development of embedded Artificial Intelligence solutions designed to optimize and revolutionize wireless communications, incorporating neuromorphic computing and satellite applications. Objectives

sites.google.com

View attachment 84766 View attachment 84767

Partnership Day 2025: Powering AI innovation

Prime Minister Luc Frieden visiting SnT's partnership Day

www.uni.lu

“AI is everywhere in our daily lives, but also in scientific research. The SnT Partnership Day 2025 showcased concrete projects.

Artificial intelligence is no longer a distant concept. It’s reshaping how we work, communicate, and innovate. On 7 May 2025, SnT’s annual Partnership Day brought together 700 participants from research, industry, and government representatives at the European Convention Centre Luxembourg (ECCL) to explore how AI is moving from theory to practice.

SnT is the University of Luxembourg’s tech research hub. The event highlighted how research partnerships are delivering real-world impact (…).”

Partnership Day 2025

Explore this photo album by SnT University of Luxembourg on Flickr!

www.flickr.com

View attachment 84771

Love it Frangipani! You go girl more excellent research in detail..I hope FF reads this, great researchers mean "plural" end of story...lawyers don't rule the narrative!..

Many are sheep ..not me, I always assess things by using my own initiative.

It's clear that you believe in Peters early work, you believe in Brainchip, but you, like me, believe in dealing with the facts as they present themselves, not fluff.

I will go on the record in saying, Antonio may have made a critical mistake in defending the Boards position with regards a US listing, clearly the Board intends working through all options, hiring professional advisors etc, but we're never leaving the ASX in our current state to France, Germany, Canada or the US...the ASX release cannot be denied, someone needs to take full responsibility for what's actually printed, own it, don't bullshit, because quite frankly, I and the rest of the Australian shareholder base aren't dense...own the mistake and don't blame Tony!

Sign some deals and all will be forgiven

Texta

 

[Bravo]

View attachment 84787

 

[Fullmoonfever]

Recently released paper funded by ESA.

Paper HERE

From what I can understand, the intent is focused on SNN vs ANN and neuromorphic processing power efficiencies etc.

They used AKD1000 for the test HW.

Appears to have come up alright but with sone work still to be refining re the SNN models.



Released 16/5.

Energy efficiency analysis of Spiking Neural Networks for space applications

Ethics declarations​

This work was founded by the European Space Agency (contract number: 4000135881/21/NL/GLC/my) in the framework of the Ariadna research program. The authors declare that they have no known competing financial interests or personal relationships that are relevant to the content of this article. The EuroSAT dataset used in this activity is publicly available at [43].

1.1 Work objectives​

With neuromorphic research being a relatively new topic, several steps are still needed to move from theoretical and numerical studies to practical implementation on real hardware flying onboard spacecraft. This work aims to contribute to this transition by providing some practical tools needed for the design of future spaceborne neuromorphic systems. The primary objective is to estimate the actual advantages that can be expected from SNN with respect to classical ANN. Here, the focus is put to the trade-off between accuracy and energy, as energy efficiency is the most prominent benefit sought in space applications. To achieve this goal, a novel metric, capable of comparing the model complexity of both ANN and SNN in a hardware-agnostic way, is proposed as a proxy for the energy consumption. The performance of several SNN and ANN models is compared on a scene classification task, using the EuroSAT RGB dataset. Special attention is placed on spiking models based on temporal coding, as they promise even greater efficiency of resulting networks, as they maximize the sparsity properties inherent in SNN, but rate-based models are included in the analysis as well. In order to validate this approach, the energy trend predicted with the proposed metric is compared with actual measures of energy used by benchmark SNN models running on neuromorphic hardware. The secondary goal of this work is to exploit the data collected in the comparison to analyze the internal dynamics of SNN models, identifying the most significant factors which affect the energy consumption, in order to derive design principles useful in future activities.

5 Conclusion​

An investigation of the potential benefits of Spiking Neural Networks for onboard AI applications in space was carried out in this work. The EuroSAT RGB dataset, a classification task representative of a class of tasks of potential interest in the field of Earth Observation, was selected as case study. SNN models based on both temporal and rate coding, and their ANN counterparts, were compared in a hardware-agnostic way in terms of accuracy and complexity by means of a novel metric, Equivalent MAC operations (EMAC) per inference. EMAC is suitable for assessing the impact of different neuron models, distinguishing the contributions of synaptic operations with respect to neuron updates, and comparing SNN with their ANN counterparts. In its base formulation, EMAC achieves dimensionless estimation, and it should then be considered only a proxy for energy consumption. Nevertheless, internal parameters can be tuned to match the features of specific hardware, if known, achieving also absolute estimation. A preliminary successful demonstration is given for the BrainChip Akida AKD1000 neuromorphic processor. Benchmark SNN models, both latency and rate based, exhibited a minimal loss in accuracy, compared with their equivalent ANN, with significantly lower (from −50 % to −80 %) EMAC per inference. An even greater energy reduction can be expected with SNN implemented on actual neuromorphic devices, with respect to standard ANN running on traditional hardware. While Surrogate Gradient proved to be an easy and effective way to achieve offline, supervised training of SNN, scaling to very deep architectures to achieve state-of-the-art performance is still an issue. Further research is needed particularly in the search of architectures capable to exploit SNN peculiarities, and in the development of regularization techniques and initialization methods suited to latency-based networks. Attention should be also given to recent developments in training techniques which do not require backward propagation in time, but only along the network at each time step [81], and new ANN-to-SNN conversion methods tailored for achieving extremely low latency [82]. Overall, the confirmed superior energy efficiency of Spiking Neural Networks is of extreme interest for applications limited in terms of power and energy, which is typical of the space environment, and SNN are a competitive candidate for achieving autonomy in space systems.