BRN Discussion Ongoing

VictorG

Member
I don't know but wouldn't be surprised if there is something for Brainchip in this, but if you feel like a little digging, I think I'll leave this here👇
Flex Logix and Brainchip have a few partners in common and their technology would work well together.

InferX™ Inference Accelerator IP​

High Throughput, Low Cost, Low Power​

The InferX AI accelerator IP is in development.
More information coming in Q1 2023.
contact us at: info at flex-logix.com
 
  • Like
  • Thinking
Reactions: 11 users
Nanose was my original reason for buying into Brainchip so it’s great to see this still in the pipeline!





1674458373085.png

 
  • Like
  • Fire
  • Love
Reactions: 37 users

wilzy123

Founding Member
Couldn't recall if posted prev and tbh didn't search.

Wonder how the update mid Dec re Transformers fits?

Is it to do with interface of third party models or to with ours or combination or...??

Might need @Diogenese thoughts if not already provided b4.


Upgrade to akida/cnn2snn 2.2.6 and akida_models 1.1.8​

Latest

Compare
@ktsiknos-brainchip
ktsiknos-brainchip released this Dec 14, 2022
2.2.6-doc-1
d334eea

Update akida and cnn2snn to version 2.2.6

New features​

  • [akida] Upgrade to quantizeml 0.0.13
  • [akida] Attention layer
  • [akida] Identify AKD500 devices
  • [engine] Move mesh scan to host library

API changes​

  • [engine] toggle_learn must be called instead of program(p,learn_enabled)
  • [engine] set_batch_size allows to preallocate inputs

Bug fixes​

  • [engine] Memory can grow indefinitely if queueing is faster than processing

Update akida_models to 1.1.8

  • updated CNN2SNN minimal required version to 2.2.6 and QuantizeML to 0.0.13
  • VWW model and training pipeline refactored and aligned with TinyML
  • Layer names in almost all models have been updated in preparation for quantization with QuantizeML
  • Tabular data models and tools have been removed from the package
  • Transformers pretrained models updated to 4-bits
  • Introduced calibration utils in training toolset
  • KWS and ImageNet training scripts now offer a "calibrate" CLI action
  • ImageNet training script will now automatically restore the best weights after training

These github commits are definitely worth watching. They are public and need to genuinely reflect changes in source code, which legitimately hold clues into the ongoings at BRN for those that know what it is they are reading.
 
  • Like
  • Fire
  • Love
Reactions: 27 users

cassip

Regular
SP in Germany is up 7,28% at Tradegate (0,4596 € / 0,7163 AUS $)
Volume low, 64k
 
  • Like
  • Fire
  • Love
Reactions: 46 users
SP in Germany is up 7,28% at Tradegate (0,4596 € / 0,7163 AUS $)
Volume low, 64k
We have shorters on the ASX
 

equanimous

Norse clairvoyant shapeshifter goddess
Food for thought

Multi-compartment Neuron and Population Encoding improved Spiking Neural Network for Deep Distributional Reinforcement Learning​

01/18/2023

by Yinqian Sun, et al.

12

share
Inspired by the information processing with binary spikes in the brain, the spiking neural networks (SNNs) exhibit significant low energy consumption and are more suitable for incorporating multi-scale biological characteristics. Spiking Neurons, as the basic information processing unit of SNNs, are often simplified in most SNNs which only consider LIF point neuron and do not take into account the multi-compartmental structural properties of biological neurons. This limits the computational and learning capabilities of SNNs. In this paper, we proposed a brain-inspired SNN-based deep distributional reinforcement learning algorithm with combination of bio-inspired multi-compartment neuron (MCN) model and population coding method. The proposed multi-compartment neuron built the structure and function of apical dendritic, basal dendritic, and somatic computing compartments to achieve the computational power close to that of biological neurons. Besides, we present an implicit fractional embedding method based on spiking neuron population encoding. We tested our model on Atari games, and the experiment results show that the performance of our model surpasses the vanilla ANN-based FQF model and ANN-SNN conversion method based Spiking-FQF models. The ablation experiments show that the proposed multi-compartment neural model and quantile fraction implicit population spike representation play an important role in realizing SNN-based deep distributional reinforcement learning.

 
  • Like
  • Love
Reactions: 15 users

Slade

Top 20
We have partnered with Renesas, Prophesee, MegaChips, Edge Impulse, Intel and ARM. As a shareholder I am more than happy with that. I have a lot of patience when it comes to Intel and ARM. I'm not stressing over the lack of a licensing fee. In fact it would not surprise me if the fee was waived for Intel and ARM. Why pay a license fee for the privilege of helping make Brainchip a huge success. Surely the royalties that roll in for years are the crux of Brainchips' business model. Watching and holding, holding and watching. I will be at the 2024 AGM and buying plenty of rounds after it.
 
Last edited:
  • Like
  • Love
  • Fire
Reactions: 62 users

GStocks123

Regular
Check this out- worth a dig through!


I notice Prophesee Gen 4 at the top of vid screen.

 
  • Like
  • Fire
  • Haha
Reactions: 21 users

GStocks123

Regular
Prophesee web page updated?

Anyone have access to the white paper?
 

Attachments

  • F3829DC9-1F0E-48D9-8731-9BA2C2079B74.jpeg
    F3829DC9-1F0E-48D9-8731-9BA2C2079B74.jpeg
    1,007.5 KB · Views: 84
  • Like
  • Thinking
  • Fire
Reactions: 18 users

GStocks123

Regular
  • Like
  • Wow
Reactions: 16 users

Slade

Top 20
  • Like
  • Love
Reactions: 15 users

Learning

Learning to the Top 🕵‍♂️
@ just 1% of the Smart Cameras market by 2027 is $350 Million, imagine the possibility.
Screenshot_20230123_214858_Chrome.jpg


OYSTER BAY, NY—Advances in machine learning technology will help propel sales of smart cameras for machine vision applications to 197 million units and a total value of $35 billion by 2027, according to global technology intelligence firm ABI Research.

“The shift from machines that can automate simple tasks to autonomous machines that can ‘see’ to optimize elements for extended periods will drive new levels of industrial innovation. This is the innovation that machine learning offers to machine vision. Machine learning can augment classic machine vision algorithms by employing the range and reach of neural network models, thus expanding machine vision far beyond visual inspection and quality control,” explains David Lobina, artificial intelligence and machine learning analyst at ABI Research.

Smart cameras, embedded sensors and powerful computers can bring machine learning analyses to every process step. Smart machine vision is already on the job in factories, warehouses and shipping centers, aiding and assisting human workers by handling mundane tasks, freeing workers to use their expertise to focus on essential processes. The market is also ripe for development in smart cities, smart healthcare, and smart transportation.

As in other cases of edge-based machine learning applications, the best way for the technology to advance is through a combination of hardware and software technology and employing information-rich data. In cases involving sensitive or private data, such as healthcare, a whole package should provide hardware (cameras, chips, etc.), software, and a way to analyze the data.

The “whole package” approach is perhaps not the most common example in the market. Still, vendors must be increasingly aware of how their offerings can mesh with other technology, often requiring hardware-agnostic software and software-agnostic data analysis.

“This is a crucial point in the case of smart cities, healthcare, and transportation, especially regarding what machine vision can achieve in all these settings. For edge vision, software and hardware vendors, as well as service providers, will start taking an expansive view of the sector,” Lobina says.


Learning 🏖 🧧🧨
 
  • Like
  • Fire
  • Love
Reactions: 32 users

FlipDollar

Never dog the boys
Our turn is coming…
 

Attachments

  • 18E12193-08D3-401E-96FE-A70F77297290.png
    18E12193-08D3-401E-96FE-A70F77297290.png
    5.9 MB · Views: 88
  • Like
  • Fire
Reactions: 13 users

charles2

Regular
SOXX (Philadelphia semiconductor index) up nearly 5% today in the US. Distinct inverse head and shoulders chart pattern since July 1. Very bullish pattern at this juncture.

If not a fake out: A rising tide will be beneficial to all semi related stocks.......even Brainchip.

(They say).
 
  • Like
  • Love
  • Fire
Reactions: 25 users
D

Deleted member 118

Guest
  • Like
Reactions: 13 users

Deadpool

hyper-efficient Ai
Obviously the consensus appears to be that we won't see much new $ flowing in this qtrly and this may turn out to be correct.

From memory there was some residual carried forward from last year which may show up you would think plus anything new would be a nice bonus.

Regardless of what we may be expecting, we know it is the wider mkt that will dictate how the SP reacts and I realistically expect the vultures to grind it further downwards if we can't show something positive trending.

What I am hopeful for is a joint release of the qtrly with maybe confirmation of say Akida 2.0 release and info re transformers etc to offset any mkt reservations with the qtrly.
Hi mate, not BRN related, but came across this music vid you may enjoy.

 
  • Haha
  • Like
  • Fire
Reactions: 10 users

equanimous

Norse clairvoyant shapeshifter goddess
At first I was confused by the name of this handle but dont think our resident wordsmith here has a twitter account.

 
  • Like
  • Haha
  • Wow
Reactions: 36 users

TopCat

Regular
I may be wrong , but I don’t remember coming across this on nViso’s website before specifically mentioning PainChek. 🤷‍♂️


29251AC9-A65D-41EE-8285-3A07FED1820A.jpeg
 
  • Like
  • Fire
  • Love
Reactions: 44 users

Steve10

Regular
Semiconductors sector in US about to golden cross on daily chart.

Nvidia +7.59% & AMD +9.22% overnight in US.

SMH_2023-01-24_08-27-54.png
 
  • Like
  • Fire
  • Love
Reactions: 43 users

buena suerte :-)

BOB Bank of Brainchip
Morning Chippers....... GREEN everywhere in US, Europe & Asia! Lets hope we follow suit!

Not long to wait now for the 4C !!! ??????????? 🤞🤞:cool:






SYMBOLPRICECHANGE%CHANGE
DJIA33,629.56+254.07+0.76
NASDAQ11,364.41+223.98+2.01
S&P 5004,019.81+47.2+1.19

Stock Indexes​

SYMBOLPRICECHANGE%CHANGE
*SHANGHAI3,264.81+24.53+0.76
*NIKKEI26,906.04UNCHUNCH
*HSI22,044.65+393.67+1.82
*SHENZHEN11,980.62+67.36+0.56
*KOSPI2,395.26+14.92+0.63
*ASX 2007,457.3UNCHUNCH
*NIFTY 5018,118.55+90.9+0.5
*STI3,293.71+17.53+0.54
*SGX-CNBC CHINA GROWTH1,473.26+0.08+0.01
*CNBC 1008,645.03+36.64+0.43
NZX 5011,974.17+25.45+0.21
*.SETI1,684.04+6.79+0.4
*MALAYSIA1,500.33+4.11+0.27
*TAIWAN14,932.93+5.92+0.0

Stock Indexes​

SYMBOLPRICECHANGE%CHANGE
*FTSE7,784.67+14.08+0.18
*DAX15,102.95+69.39+0.46
*CAC7,032.02+36.03+0.52
*STOXX600454.49+2.37+0.52
*AEX747.87+9.23+1.25
*BEL 203,889.84+28.39+0.74
*FTSE MIB25,821.45+45.93+0.18
OMXS302,224.53+27.11+1.23
*SMI11,406.27+111.25+0.98
HEX11,232.18+80.2+0.72
*PSI205,972.56+58.75+0.99
OMXC 251,729.49+6.63+0.38
*RUSSIA2,185.31+18.62+0.86






 
  • Like
  • Love
  • Fire
Reactions: 37 users
Top Bottom