Last week, HeCaiJing (“核财经”), a prominent Chinese blockchain media outlet, interviewed Matrix Chief AI Scientist Professor Steve Deng. In the interview, Professor Deng discussed several topics including the recently revealed YangDong Artificial Intelligence Research Institute and Matrix’s Enterprise-Grade Big Data Cloudchain Solution — tentatively the Matrix Enterprise Cloudchain, for short. The Matrix Enterprise Cloudchain is a high-performance, secure and interoperable enterprise-grade big data platform combining AI, cloud computing, and blockchain technology.
This article serves to share some highlights from the interview with Matrix’s non-Chinese-speaking community. Please note that this is not a direct translation of the original interview. Rather, choice quotes and comments are pulled and given context. The original interview can be found here.
On the Matrix AI Network
The Matrix AI Network aims to tackle two big problems slowing the development of AI. Namely, the lack of computing power available for AI research and the relative lack of verified data necessary to produce qualitative improvement in AI models for social good. The Matrix Enterprise Cloudchain is designed not only to promote the further development of the Matrix AI Network ecosystem, but also to promote the integration of both AI and blockchain technology in everyday life to create real economic value.
On Decentralizing AI
Strictly speaking, AI has historically relied on centralized research data. According to Professor Deng, Matrix combines AI with the decentralized and distributed characteristics of blockchain technology to enable several key components on the Matrix Enterprise Cloudchain. First, using the Matrix AI Network, computing power from several distinct sources can be integrated into the Enterprise Cloudchain platform to power applications beyond the scope of standard personal equipment. Second, the Matrix AI Network allows the Enterprise Cloudchain to safely store data due to, amongst other things, Matrix’ IPFS support. Finally, and perhaps most importantly, Matrix’s blockchain technology helps solve issues of data validation.
According to Professor Deng, “many enterprises have a lot of data, but they are not willing to give you access to their data to train AI models because they have fears about data privacy and improper data disclosure. Besides these fears, enterprises also worry that once the data has been used to train the AI model, the AI model might be used to generate revenue. They worry that they won’t get a fair share of this revenue.”
Professor Deng believes that the solution to this problem is found in blockchain technology. Blockchain technology makes it possible to guarantee the rights and interests of everyone involved — be it the data providers, AI-model developers and users. With the kind of protection that blockchain provides, major obstacles slowing the development of AI can be overcome. “Data and AI models are inherently digital, so we aim to take into account the interests of all parties using our enterprise-grade big data cloudchain solution,” reiterates Professor Deng.
On AI and Blockchain
From a technical point of view, although blockchain and artificial intelligence are two distinct technologies, Professor Deng believes that the integration of both is necessary to maximize their potentials. “The integration of AI and blockchain technology brings new possibilities.” He points out that, in the near future, the integration of data and AI models will bring the first of many great achievements made possible due to the provision of computing power for data validation and AI model training over the blockchain. This will form the foundation of an open, credible AI platform for Internet of Things applications.
On Industrial Big Data
Industrial Big Data is at the intersection of big data, the Internet of Things, and industrial equipment. Industrial Big Data is also the central peg of several Chinese national strategic plans including the “Action Plan for the Development of Industrial Internet”, “Made in China 2025” and “Industry 4.0”.
While domestic in nature, these Chinese national strategies have far reaching implications as China is a major player in industrial equipment manufacturing, operations and maintenance industries. For instance, China is responsible for about 80% of the world’s total port equipment, 70% of the world’s high-speed trains, 60% of the world’s excavators, 40% of the world’s marine vessels and 39% of the world’s heavy equipment. However, despite these lofty achievements, inefficiencies abound.
According to recent data, while China created 11.6% of the world’s GDP in 2013, it did so while accounting for 21.3% of the world’s energy consumption. According to Professor Deng, these industries are primed for a revolution. “When only considering the rail industry, a quarter of the world’s transport workload happens in China despite using only 6% of the world’s railway operating mileage,” notes Deng. “This is the highest transport density in the world.”
Unlike traditional internet applications, industrial big data applications tend to have higher and more stringent requirements for real-time processing. As network bandwidth is often limited, it can be difficult to guarantee performance by solely relying on cloud platforms. Similarly, industrial big data has stringent privacy requirements. This means that using existing public cloud solutions is not appropriate. Professor Deng notes that “the cost required for an enterprise to build, operate and maintain a private cloud is prohibitive. It is not a cost-effective or viable option for most enterprises.”
On Equipment Manufacturing, Operations and Maintenance
In this blockchain era, large industrial data platforms like the Matrix Enterprise Cloudchain have the potential to create value across industries and countries. Professor Deng notes that the technology underlying Matrix’s Enterprise Cloudchain is already being used and tested in an overseas locomotive remote monitoring system servicing Kenya’s Mombasa–Nairobi Standard Gauge Railway. The system is monitoring 48 locomotives exported to Kenya by the CRRC Corporation Limited. Professor Deng also indicates that the system is supported by the National Natural Science Foundation of China, the CRRC and the China Railway Corporation, amongst others.
On Medical Treatments
Professor Deng goes on to reiterate that Matrix is engaged with several hospitals to deploy and use Matrix’s cancer diagnosis services. He posits that “the uniqueness of this system is that, first, we pay more attention to the challenging issues based on the obtained diagnostic and therapeutic data and laboratory test results. Second, we are cooperating with leading hospitals in the industry to tack frontier issues. Third, we have high-quality patient data from early screening to CT imaging to biopsy. So far, we have more than 2000 patients’ data — all of which are labeled by top Chinese doctors.” This type of data is invaluable to the development and training of AI models.
On Remaining Barriers
As Professor Deng points out, “manufacturing industries — especially high-end manufacturing industries — form the foundation of a country. While these industries create value for society and provide employment opportunities, profits themselves are actually quite low.” Matrix aims to help the continued development of these industries by reducing inefficiencies using its Enterprise Cloudchain. Additionally, the Matrix AI Network remains keenly aware of the importance of hardware. “At present, the barrier to the amalgamation of AI and blockchain lies in hardware development.” Professor Deng says that the Matrix AI Network continues to ramp up R&D on its proprietary mining machine chip. “The idea is that these chips are specially designed to efficiently run deep learning algorithms and AI models.” A hardware solution lowers the risk of algorithm and model tampering and reduces the reliance on CPUs and GPUs.
On the Future
Professor Deng shares that, “presently, we hope to increase cooperation with local-level governments to establish research institutes and centers, especially in more developed manufacturing parks.” With the ongoing development of science and technology, artificial intelligence is poised to change our daily lives. Matrix hopes to connect people and business applications around the world to power AI for social good.