NetMind.AI——Decentralized AI at the core of Tsinghua University

Original title: "Tsinghua team launches new platform: using decentralized AI to break the computing power shortage"

Original source: Quantum Bit

Recently, a piece of data pointed out the astonishing growth in the demand for computing power in the field of AI. According to estimates by industry experts, Sora launched by OpenAI needs to be trained on 4200-10500 NVIDIA H100s for about 1 month in the training phase, and when the model is generated to the inference phase, the computing cost will quickly exceed the training phase.

If this trend continues, the supply of GPUs may be difficult to meet the continuous demand for large models.

However, there has been a new trend overseas recently, which may provide a new solution to the upcoming "computing power shortage" - decentralized AI.

Three weeks ago, on March 23, Stability AI suddenly issued an announcement announcing the resignation of the company's CEO Emad Mostaque. Emad Mostaque himself revealed the next move, to pursue the "dream of decentralized AI".

However, due to the technical pain points such as uncertainty and instability of decentralized networks, it is difficult for the previous wave of decentralized AI to truly land in the era of large models.

Until recently, Quantum Bit discovered that a Tsinghua team that started a business overseas focused on decentralized AI and founded NetMind.AI. In 2023, NetMind released a white paper detailing the decentralized computing power sharing platform NetMind Power. This platform aims to solve the pain points of decentralized AI landing in the era of large models.

1. Make GPU affordable for every developer

In September 2021, NetMind.AI launched a decentralized computing platform project called NetMind Power.

There is a large amount of idle computing power in the world: idle computing power in traditional data centers, underutilized computing power owned by small and medium-sized enterprises, and scattered GPUs owned by individuals. These computing powers are either idle or used for gaming and video rendering. At the same time, AI computing power is becoming increasingly scarce. AI researchers, small and medium-sized enterprises, especially AI startups, and traditional companies involved in AI projects are all trapped by the high cost and high threshold of AI computing power.

So, NetMind Power created a decentralized computing network, using the core technology developed by NetMInd to leverage global computing resources and provide easy-to-use and affordable AI computing services for the AI industry.

△NetMind Power is an economical choice for obtaining computing power, providing users with efficient and affordable computing resource solutions.

Currently, NetMind Power has collected thousands of graphics cards, including H100, A100, 4090, and 3090.

Four highlights of the platform:

1. Decentralized dynamic cluster - Building reliable and efficient AI applications on extremely uncertain computing power

The Power platform uses P2P-based dynamic distributed cluster technology, combined with its unique routing, clustering algorithms and neural networks, to weave thousands of computing nodes into a powerful network cluster architecture, specifically serving high-level needs such as AI applications.

When users perform AI-related operations on the Power platform, such as model training, fine-tuning or reasoning, Power's decentralized network can quickly allocate the most appropriate computing resources in computing nodes around the world through optimization algorithms in a very short time to provide services to users.

At the same time, Power provides B-side users with dynamic cluster strategies, which can intelligently reorganize and configure nodes within a few seconds, and provide customizable, highly scalable and highly redundant exclusive clusters.

2. Complete AI Ecosystem: Lower the threshold for computing power usage and expand decentralized network application scenarios

With NetMind's years of accumulation in the field of AI, Power Network will include AI ecological foundations such as open source model libraries, AI data sets, data and model encryption, as well as full-service such as model training, reasoning, and deployment in addition to basic computing power services, to create a MaaS (Model as a Service) platform to empower both computing power suppliers and AI application parties.

For AI+ projects of scientific researchers, SMEs in the AI field, and traditional enterprises, Power's MaaS platform will significantly lower the threshold for computing power usage, especially for SMEs and traditional enterprises without professional AI development capabilities. This is particularly important.

For traditional computing power suppliers, the Power network can reach more users. Furthermore, with the help of Power's MaaS platform, they can expand application scenarios and obtain higher returns. In this way, the Power network can also incorporate traditional small and medium-sized centralized computing power into the decentralized computing power network, thereby greatly expanding the scale of the network.

3. Asynchronous training algorithm - Solve network bottlenecks and tap the potential of idle computing power

In the current field of machine learning, especially in the training of large-scale language models, it is usually necessary to use GPU dedicated connection lines or high-bandwidth internal networks to achieve synchronous distributed training between GPUs, which inevitably increases the threshold and cost of training.

NetMind Power breaks the barriers of network speed and bandwidth in distributed training through self-developed model segmentation and data asynchrony technology. Even training nodes distributed in different corners of the earth can simultaneously participate in the huge model training work.

4. Model encryption and data isolation - Solving security problems in decentralized networks

Power provides unique model encryption technology to ensure the security of users' AI models and data in decentralized volunteer computing scenarios. All network communications are encrypted to ensure the security of data transmission; through data isolation and model splitting, it is ensured that no single node in the decentralized network can obtain complete data and models, greatly improving security.

Second, another Tsinghua-background team that has been entrepreneurship overseas for many years

The core team of NetMind.AI comes from Tsinghua and has been honing in the field of AI for more than 10 years.

The company's founder and CEO, Kai Zou, graduated from Tsinghua University's Basic Sciences in Mathematics and Physics in 2010 and received a master's degree in mathematics and statistics from Georgetown University in 2013.

He is a serial entrepreneur who has led both ProtagoLabs and the nonprofit AGI Odyssey. He is also an angel investor who has invested in several AI startups including Haiper.ai, Auto Edge, Qdot, and Orbit.

It is worth noting that the paper "EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks" published by Kai Zou and OpenAI researcher Jason Wei has been cited more than 2,000 times. The CEO and his team firmly believe that the platform they have built should provide resources for scholars who are truly engaged in academic research and corporate engineers who are promoting the development of AI.

The company's CTO received a master's degree in computer science from George Washington University in 2016. Before joining NetMind.AI, he served as a senior team leader at Microsoft. He has extensive experience in Web3, blockchain technology, distributed systems, Kubernetes, cloud computing, Azure and AWS, and has professional skills in edge computing, full-stack development and machine learning.

Third, the ultimate ideal: bringing AI to thousands of households

Behind NetMind's vision of decentralized AI, there is actually a deep ideal of technology inclusion.

Looking back at the history of IT technology development, the trend of decentralization often emerges at a time when computing resources are concentrated sharply. As a bottom-up force, it fights against giants who try to monopolize all resources, thus opening a new round of technology inclusion and allowing new technologies to truly spread to every corner of the world.

Today's large model market may be at such a moment.

Looking at the large model market, after a year of vigorous development, there are not many startups that can really gain a foothold. Except for a few star unicorns, the future of large models seems to be converging in the hands of technology giants such as Microsoft, Google, and Nvidia. In the long run, a few companies may form a monopoly control over the pricing, availability, and access rights of computing resources.

At this time, a democratic narrative like NetMind Power is needed to write a new blueprint for the story of AGI.

Currently, NetMind has cooperated in academic and commercial fields -

In terms of academics, NetMind Power has currently cooperated with many top domestic and foreign universities, including Cambridge University, Oxford University, Carnegie Mellon University, Northwestern University, Tsinghua University, Huazhong University of Science and Technology, Rice University, Fudan University, Shanghai Jiaotong University, etc.

In terms of business, NetMind Power provides AI computing solutions based on decentralized networks, allowing enterprises to focus on model development and product innovation. More and more companies are accelerating the launch of innovative AI products with the help of Netmind Power. For example, Haiper.ai, a Wensheng video team that has been gaining momentum in North America recently, has deeply integrated its model training and reasoning with the NetMind Power platform.

In the future, NetMind Power will gradually grow into a decentralized AI community to accelerate global AI innovation.

Machine learning practitioners, academic researchers, and AI application companies can find the computing power and models they need on the NetMind Power platform, and can also host their trained models on the platform, or even provide them to other users on the platform and charge a certain fee.

Users can not only call the corresponding computing power on the platform to solve their training needs, but also provide their trained models to more people or companies in need through the platform, passing them on layer by layer.

To extend the timeline, to truly realize AGI, the universalization and democratization of AI are inevitable prerequisites. Today, NetMind.AI, which took the lead, is making its own contribution, looking for more partners, and taking a solid step towards the democratic AGI era.

This article is from a contribution and does not represent the views of BlockBeats.

Original link

欢迎加入律动 BlockBeats 官方社群：

Telegram 订阅群： https://t.me/theblockbeats

Telegram 交流群： https://t.me/BlockBeats_App

Twitter 官方账号： https://twitter.com/BlockBeatsAsia