HopalongPetrovski
I'm Spartacus!
Her Dad must be so proud.Fantastic PR and advice on how to increase the Brainchip share price
Someone can send this to Sean & the team
Have a great weekend everyone!
View attachment 65615
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Humble Genius
Regular
Off course he is - she owns shares in BRCHFHer Dad must be so proud.
Hi Bravo,Isn't this similar to what we're trying to achieve to slash power consumption @Diogenese?
SNN's eliminating the need for matrix multiplication without affecting performance?
Researchers claim new approach to LLMs could slash AI power consumption
Work was conducted as concerns about power demands of AI have risen
Graeme Burton
27 June 2024ā¢ 2 min read
SHARE
ChatGPT uses 10 times the electricity of a Google search ā but cutting out āmatrix multiplicationā could cut this number without affecting performance
New research suggests that eliminating the āmatrix multiplication' stage of large-language models (LLMs) used in AI could slash power consumption without affecting performance.
The research was published at the same time that increasing concerns are being raised over the power demands that AI will make over the course of the decade.
Matrix multiplication ā abbreviated to MatMul in the research paper ā performs large numbers of multiplication operations in parallel, becoming the dominant operation in the neural networks that drive AI primarily due to GPUs being optimised for such operations.
But by leveraging Nvidia CUDA [Compute Unified Device Architecture] technology instead, along with optimised linear algebra libraries, this process can be efficiently parallelised and accelerated ā cutting power consumption without penalising performance.
In the process, the researchers claim to have "demonstrated the feasibility and effectiveness of the first scalable MatMul-free language model", in a paper entitled Scalable MatMul-free Language Modelling.
The researchers created a custom 2.7 billion parameter model without using matrix multiplication, and found that performance was on a par with state-of-the-art deep learning models (called ātransformers', a fundamental element of natural language processing). The performance gap between their model and conventional approaches narrows as the MatMul-free scales, they claim.
They add: "We also provide a GPU-efficient implementation of this model, which reduces memory usage by up to 61% over an unoptimised baseline during training. By utilising an optimised kernel during inference, our model's memory consumption can be reduced by more than 10 times compared to unoptimised models."
Less hardware-heavy AI models could also enable more pervasive AI, freeing the technology from its dependence on the data centre and the cloud. In addition, both OpenAI and Meta are set to unveil new models that, they claim, will be capable of reasoning and planning.
However, they cautioned, "one limitation of our work is that the MatMul-free language model has not been tested on extremely large-scale models (for example, 100+ billion parameters) due to computational constraints". They called for well-resourced institutions and organisations to build LLMs utilising lightweight models, "prioritising the development and deployment of matrix multiplication-free architectures".
Moreover, the researchers' work has not yet been subject to peer review. Indeed, power consumption on its own matters less than energy usage per unit of āoutput' ā information that is not provided in the research paper.
Researchers claim new approach to LLMs could slash AI power consumption
ChatGPT uses 10 times the electricity of a Google search ā but cutting out āmatrix multiplicationā could cut this number without affecting performancewww.computing.co.uk
Scalable MatMul-free Language Modeling
Matrix multiplication (MatMul) typically dominates the overall computational cost of large language models (LLMs). This cost only grows as LLMs scale to larger embedding dimensions and context lengths. In this work, we show that MatMul operations can be completely eliminated from LLMs while...arxiv.org
This reads like a von Neumann computer software implementation, not an SNN. They run it on a GPU.
The software makes it slow and power hungry compared to Akida.
It may be ok for cloud processing, but may not be suitable for real-time applications.
CHIPS
Regular
I think Kuchenbuch urgently needs to work on his presentation skills. There are courses for that.
Already his first slide and and confusing presentation made me shut it off.
Fullmoonfever
Top 20
Looks like Pedro is enjoying his new toy so far
LinkedIn Login, Sign in | LinkedIn
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.www.linkedin.com
Pedro Machado, a Senior Lecturer in Computer Science at Nottingham Trent University must have been a little giddy and trembling with excitement while typing this LinkedIn post about his uni joining āthe race of Neuromorphic Computing by being [one] of the pioneers accelerating SNNs on the Alkidaās Brainchip.ā
View attachment 64801
View attachment 64803
View attachment 64804
Akida will be useful for their ongoing project on āPosition Aware Activity Recognitionā:
View attachment 64805
Search Ć¢ā¬Ā¢ Threads
www.threads.net
Agree. 25 years of experience as a leading salesperson. I think shareholders can demand a better presentation than that.
Bring Rob Telson back!
We want the share price to go lower so we can buy back in. Hehehe
CHIPS
Regular
We want the share price to go lower so we can buy back in. Hehehe
You had so many chances to buy low and did not use them? And now you want the SP to fall again after it has recovered a bit?
How stupid can one be?
sarcasm....................
I am a stubbie short of a six pack.
Tothemoon24
Top 20
Abstract
Analysing a visual scene by inferring the configuration of a generative model is widely considered the most flexible and generalizable approach to scene understanding. Yet, one major problem is the computational challenge of the inference procedure, involving a combinatorial search across object identities and poses. Here we propose a neuromorphic solution exploiting three key concepts: (1) a computational framework based on vector symbolic architectures (VSAs) with complex-valued vectors, (2) the design of hierarchical resonator networks to factorize the non-commutative transforms translation and rotation in visual scenes and (3) the design of a multi-compartment spiking phasor neuron model for implementing complex-valued resonator networks on neuromorphic hardware. The VSA framework uses vector binding operations to form a generative image model in which binding acts as the equivariant operation for geometric transformations. A scene can therefore be described as a sum of vector products, which can then be efficiently factorized by a resonator network to infer objects and their poses. The hierarchical resonator network features a partitioned architecture in which vector binding is equivariant for horizontal and vertical translation within one partition and for rotation and scaling within the other partition. The spiking neuron model allows mapping the resonator network onto efficient and low-power neuromorphic hardware. Our approach is demonstrated on synthetic scenes composed of simple two-dimensional shapes undergoing rigid geometric transformations and colour changes. A companion paper demonstrates the same approach in real-world application scenarios for machine vision and robotics.Getupthere
Regular
Samsung backs āworldās most powerfulā AI chip for edge devices ā The Next Web
Eindhoven-based startup Axelera has raised $68mn as it looks to take its AI chip business global. One of the lead investors is Samsung Catalyst, the venture arm of the semiconductor giant Samsung Electronics. Axelera is developing AI chips that enable computer vision and generative AI in edge...
Gies
Regular
BrainChip on LinkedIn: #ai #neuromorphic
Have you downloaded our recent white paper on TENNs-PLEIADES (Polynomial Expansion In Adaptive Distributed Event-based Systems)? Unlike traditional methodsā¦
Pom down under
Top 20
Pom down under
Top 20
Pom down under
Top 20
Guess they are planning using SNN for sometime and I like the last paragraph bit at the bottom 
The purpose of this internship is to implement low level event processing using SNN accelerators.
www.linkedin.com
PROPHESEE
Founded by the worldās leading pioneers in the field of neuromorphic vision, Prophesee develops computer vision sensors and systems for application in all fields of artificial vision. The sensor technology is inspired by biological eyes, acquiring visual information in an extremely performing yet highly efficient way. Propheseeās disruptive vision sensor technology entirely overthrows the established paradigms of frame-based vision acquisition currently used everywhere in computer vision.
This is a great opportunity to join a dynamic company and an exciting team and to lead a paradigm shift in computer vision across many industries.
EVENT-BASED TECHNOLOGY
Prophesee designs and produces a new type of cameras that are bio-inspired and thus free themselves from the concept of images. They do not gather information with a fixed framerate but instead each pixel is captured asynchronously when needed. These bio-inspired cameras, also called event-based cameras or neuromorphic cameras, therefore have an extremely sparse output and enable, with appropriate algorithms, a real time treatment of the information at an equivalent frequency of a kHz or more. But since the data coming from the sensor are quite different from the images traditionally used in standard vision, Prophesee is also advancing the algorithmic and machine learning side of this new kind of machine vision. This enables our clients to build new applications in a wide range of domains, including industrial automation, connected devices, autonomous vehicles, augmented and virtual reality, and more.
Internship Position Description
We are looking for passionate interns who demonstrate initiative, take ownership for project work, and exhibit a strong spirit of innovation. The ideal candidate is a curious and creative individual keen on problem-solving and with prior experience in C++ / Python programming and exposure to the Computer Vision / Image Processing / Artificial Intelligence / Machine Learning domains.
She/He will work in a mixed team of scientists and engineers to design, develop & optimize solutions to research problems. Her/His main contribution will be in creating innovative bio-inspired computer vision algorithms for specific tasks across many applications such as computational imaging, 3d sensing, robotics, localization, factory automation, smart devices, aerospace & defense, automotive, etc.
The main required skills common to MOST internship positions are the following (but not exclusively):
Event Signal Processing versus Spiking Accelerators
Event sensors generate sparse and asynchrounous data which are not compatible with Van Neuran conventionnal computers: the states or memory of any event driven algorithm is tight with the computing part. To scale with larger pixel arrays, the bandwith of these sensors have increased and thus the same applies with the need of low level filtering close to the pixels. Many hardware accelerators have been proposed in the state of the art to ease the event processing, either using FPGAs or dedicated ASICs.
Convolutionnal neural networks are mostly sparse after few layers, and some hardware accelerators already use this feature to faster computations. However, the input information size and frequency has to be defined and fixed. Spiking Neural Networks (SNN), which could be seen as asynchronous recurrent neural networks, are adding asynchronous time based processing to traditionnal neural networks. This feature makes them suitable for event-based data, and the ability to program SSN accelerators will enable application-specific filtering.
The purpose of this internship is to implement low level event processing using SNN accelerators.
The Plan Is
What do they mean or looking for when they say this in the above?
Part of this work has to make sure that an existing accelerator can scale up.
The purpose of this internship is to implement low level event processing using SNN accelerators.
LinkedIn Login, Sign in | LinkedIn
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.
Job description
Internship opportunities at PropheseePROPHESEE
Founded by the worldās leading pioneers in the field of neuromorphic vision, Prophesee develops computer vision sensors and systems for application in all fields of artificial vision. The sensor technology is inspired by biological eyes, acquiring visual information in an extremely performing yet highly efficient way. Propheseeās disruptive vision sensor technology entirely overthrows the established paradigms of frame-based vision acquisition currently used everywhere in computer vision.
This is a great opportunity to join a dynamic company and an exciting team and to lead a paradigm shift in computer vision across many industries.
EVENT-BASED TECHNOLOGY
Prophesee designs and produces a new type of cameras that are bio-inspired and thus free themselves from the concept of images. They do not gather information with a fixed framerate but instead each pixel is captured asynchronously when needed. These bio-inspired cameras, also called event-based cameras or neuromorphic cameras, therefore have an extremely sparse output and enable, with appropriate algorithms, a real time treatment of the information at an equivalent frequency of a kHz or more. But since the data coming from the sensor are quite different from the images traditionally used in standard vision, Prophesee is also advancing the algorithmic and machine learning side of this new kind of machine vision. This enables our clients to build new applications in a wide range of domains, including industrial automation, connected devices, autonomous vehicles, augmented and virtual reality, and more.
Internship Position Description
We are looking for passionate interns who demonstrate initiative, take ownership for project work, and exhibit a strong spirit of innovation. The ideal candidate is a curious and creative individual keen on problem-solving and with prior experience in C++ / Python programming and exposure to the Computer Vision / Image Processing / Artificial Intelligence / Machine Learning domains.
She/He will work in a mixed team of scientists and engineers to design, develop & optimize solutions to research problems. Her/His main contribution will be in creating innovative bio-inspired computer vision algorithms for specific tasks across many applications such as computational imaging, 3d sensing, robotics, localization, factory automation, smart devices, aerospace & defense, automotive, etc.
The main required skills common to MOST internship positions are the following (but not exclusively):
- Excellent programming skills in C++ and/or in Python
- Engineering background in Computer Science, Mathematics or related field
- Prior experience in projects involving at least one of the following domains:
- Algorithmic design, e.g. 3d vision, machine learning, numerical optimization, etc
- Software development, e.g. implementation, architecture, optimization, testing, porting on embedded platforms, etc
- Development operations, e.g. source code versioning, continuous integration & deployment, cloud computing, system administration, etc
Event Signal Processing versus Spiking Accelerators
Event sensors generate sparse and asynchrounous data which are not compatible with Van Neuran conventionnal computers: the states or memory of any event driven algorithm is tight with the computing part. To scale with larger pixel arrays, the bandwith of these sensors have increased and thus the same applies with the need of low level filtering close to the pixels. Many hardware accelerators have been proposed in the state of the art to ease the event processing, either using FPGAs or dedicated ASICs.
Convolutionnal neural networks are mostly sparse after few layers, and some hardware accelerators already use this feature to faster computations. However, the input information size and frequency has to be defined and fixed. Spiking Neural Networks (SNN), which could be seen as asynchronous recurrent neural networks, are adding asynchronous time based processing to traditionnal neural networks. This feature makes them suitable for event-based data, and the ability to program SSN accelerators will enable application-specific filtering.
The purpose of this internship is to implement low level event processing using SNN accelerators.
The Plan Is
- SNN accelerator/processing state of the art analysis: Most of the proposed architectures are not suitable with the bandwith of event sensors. Part of this work has to make sure that an existing accelerator can scale up.
- Software Imlementation of low-level processing functions to filter the event stream using SNN. The algorithms will be similar to the filters implemented inside the event signal processor of Prophesee sensors or will be adapted from recently published academics works.
- Hardware implementation using a SNN accelerator from Prophesee partners.
What do they mean or looking for when they say this in the above?
Part of this work has to make sure that an existing accelerator can scale up.
Last edited:
Pom down under
Top 20
Prophesee delivers Event-based AMD Kria Starter Kit
Developers can now take full advantage of Prophesee Event-based MetavisionĀ® sensor and AI performance, power, and speed to create the next generation of Edge AI machine vision applications running on AMD platforms.
www.prophesee.ai
PROPHESEE: Event-based sensors for industrial cameras | AMD posted on the topic | LinkedIn
PROPHESEE's event-based image sensors, powered by AMD adaptive SoCs, reveal what traditional industrial cameras in the factory can't, improving productivityā¦
Last edited:
SharesForBrekky
Regular
For anyone interested, here is the latest Edge AI box by NVIDIA/Intel:
www.cnx-software.com
"DFI X6-MTH-ORN is a fanless Edge AI Box Computer that combines an NVIDIA Jetson Orin NX/Nano AI module with a 14th Gen Intel Core Ultra āMeteor Lake-Uā 15W processor for AI-driven applications leveraging GPU computing, machine learning, and image processing."
As compared to Brainchip's Edge AI box https://shop.brainchipinc.com/products/akidaā¢-edge-ai-box
DFI X6-MTH-ORN fanless Edge AI Box computer combines 14th Gen Intel Core Ultra SoC with NVIDIA Jetson Orin AI module - CNX Software
DFI X6-MTH-ORN is a fanless Edge AI Box Computer that combines an NVIDIA Jetson Orin NX/Nano AI module with a 14th Gen Intel Core Ultra "Meteor Lake-U"
www.cnx-software.com
"DFI X6-MTH-ORN is a fanless Edge AI Box Computer that combines an NVIDIA Jetson Orin NX/Nano AI module with a 14th Gen Intel Core Ultra āMeteor Lake-Uā 15W processor for AI-driven applications leveraging GPU computing, machine learning, and image processing."
As compared to Brainchip's Edge AI box https://shop.brainchipinc.com/products/akidaā¢-edge-ai-box
Similar threads
