Brain-like machine learning tools
Our initiative, eCortechs AB, is focused on commercializing BCPNN, a brain-compatible machine-learning technology. Diverging from traditional AI, it learns mostly without labels and is hardware-friendly, scalable, and exhibits superior performance against unseen test noise.
We want to complement current mainstream machine learning technology with BCPNN technology. Therefore, our mission is to demonstrate the superiority of BCPNN through various industry collaborations.
High-tech industries with in-house AI labs and projects.
Frustration with mainstream AI and machine learning technology, which is hard to use in practice due to need for large amounts of labeled training samples and provide non-robust and unreliable solutions.
It's a serious problem that the technology often does not live up to expectations and substantial investments fail to deliver results.
BCPNN technology can train on few labled training samples and many unlabeled samples. BCPNN-based solutions often generalize well and are robust.
It runs on GPU technology with potential to be implemented with FPGA-acceleration and ASIC.
BCPNN uses Bayesian-Hebbian learning rules and data-driven structural plasticity. Utilizes readily available unlabeled data to improve internal representations.
To quickly develop efficient and robust solutions to their machine learning problems.
We estimate the market size for which our solution is designed in monetary terms as follows
Is our goal in the next 3 years
My key role is Founder, Research lead, CTO and I'm responsible for setting strategy, conducting research initiatives & overseeing technical operations.
3 M.S.c in computer science, 3 P.h.D in computer science
Potential competitors could include high-tech companies focused on AI solutions, such as Apple and Facebook's Meta Platforms. Furthermore, any start-ups or established firms working on innovative machine-learning models can also be seen as competition.
It takes an intuitive approach to machine learning problems and can use available non-labeled and labeled data to achieve a working solution.
They pay us for a unique set of brain-like machine learning tools and a deep designer competence to use this tools for practical problems
A new more optimized version of the framework has been developed and the first FPGA-implementation has been made. A first real-world application with company partners is in progress.
Recruited a handful of application engineers, developed professional tools and FPGA implementations and achieved several success stories.
We have not launched our product yet.
Our company incorporated in
Sweden
Fast recruitment and onboarding of skilled employees
$
200000
We raised investments
Private investors with plenty of money
$
1000000
Currently, we are raising investments
$
10000000
Estimated pre-investment valuation of the company
