London, UK, April 9, 2024, Chainwire
NeuroMesh (nmesh.io), a pioneer in artificial intelligence, announces the launch of a distributed AI training protocol ready to revolutionize global access and collaboration in AI development. Embracing DePIN’s decentralized framework, NeuroMesh bridges the gap between the needs for large-scale AI model training and distributed GPUs. This initiative aims to strengthen the inclusivity of AI development and foster participation across different sectors and geographies.
Vision in AI: The team’s global ambitions
The NeuroMesh team, comprised of researchers and engineers from Oxford, NTU, PKU, THU, HKU, Google and Meta, pioneers a democratic AI training process. This visionary approach addresses the limitations of centralized AI development by enabling GPU owners around the world to contribute to a broad training network and enabling businesses of all sizes to leverage this service for their training needs.
NeuroMesh goes beyond traditional AI by fostering collaboration. Their vision is to equip every developer and organization, regardless of location or resources, with the ability to train and utilize cutting-edge AI models. This aligns perfectly with the vision of AI pioneers like Yann LeCun, who advocate for a future powered by crowdsourcing and distributed AI training.
Innovative design based on PCN
At the heart of NeuroMesh’s distributed training protocol is the groundbreaking Predictive Coding Network (PCN) training algorithm, a true game-changer in the field. This approach allows GPU owners around the world to unleash their capabilities and foster broader collaborative efforts.
PCN Training Algorithm: The magic of NeuroMesh lies in its PCN training algorithm. Unlike traditional backpropagation (BP) methods, PCN enables fully local, parallel, and unsupervised training. The team aims to create a massive network where each node representing a participating GPU learns independently. PCN minimizes communication between layers, reducing data traffic and promoting asynchronous training. Think of it as a symphony, with each musician playing his or her part independently but contributing to a harmonious whole.
Inspired by recent advances in neuroscience research pioneered at the University of Oxford, this cutting-edge model mimics the local learning approach of the human brain. We replicate the behavior of brain neurons by storing error values and optimizing for local targets in each layer. This allows NeuroMesh to define much larger models with individual components contributing to the same ultimate optimization goal for the overall network, much like the human brain where different stimuli are processed by different groups of neurons.
This biologically inspired approach, combined with unique deployment capabilities, opens a new era in AI development.
Request to build a global partnership
NeuroMesh invites partnerships globally with the goal of building an AI future where everyone can participate. NeuroMesh’s protocol is the foundation on which a diverse ecosystem is built. The ecosystem is designed to be dynamic, collaborative, and adaptable to meet the training needs of AI models in any industry and at any scale.
Individuals, projects with GPU resources, and organizations in need of training can all participate in this innovative initiative. For comprehensive details about NeuroMesh and to participate in this cutting-edge effort, visit nmesh.io.
Introducing NeuroMesh
NeuroMesh is made up of researchers and engineers from respected institutions such as Oxford, NTU, PKU, THU, HKU, Google and Meta. By helping developers and organizations deploy powerful AI models, NeuroMesh is fostering a comprehensive AI ecosystem, bridging the gap between large-scale AI model training needs and globally distributed GPUs.
Visit NeuroMesh to learn more. Twitter | telegram
contact
CMO
Kencia Lee
Neuromesh
kenchia@nmesh.io
07746906341