NuNet will enable the connection of various devices and agents in a secure and decentralized manner.
Below are potential use cases for the NuNet platform and Industry 4.0
Flexible decentralized computations at the edge
NuNet leverages the computing frameworks of its partners by allowing to build flexible and radically decentralized computation graphs spanning IoT devices of different capacities and owned by different economic players and community members - simple or advanced sensors, robotic microcontrollers, embedded systems, virtual machines on the edge, fog and cloud. It enables to the design of efficient and fast data and AI workflows for dynamic IoT environments where huge amounts of streaming data can be processed as close to the edge as required by the business model and capacities of the particular system. In the future, NuNet and SingularityNET are planning to partner for implementing technologies required for the automatic adaptability for balancing computing loads in IoT networks in real-time.
Decentralized AI model ensembles
Training current state of the art machine learning models involving massive datasets is prohibitively expensive -- e.g. the specifically designed supercomputer for training OpenAI’s costs over 250M US dollars24, while the training of the model itself costs millions of dollars. Therefore, the development and application of cutting edge AI and ML technology for most researchers, individuals and SMEs is not affordable.
Researching, training, and using AI and ML would become much more affordable and beneficial for society and the economy if latent cheap computing power of small personal devices, PCs, and possibly mobile phones could be utilized to spread the computational work required to train such models. The underlying problem historically has been addressed by volunteer computing frameworks (e.g. BOINC) and projects (e.g. Folding@Home). However, most of history and to some extent recent volunteer or distributed computing frameworks are highly application-specific, and also logically and/or architecturally centralized in the sense that there is a single server, node, or process which is responsible for the network’s operation, distribution of workloads and aggregating results.
NuNet framework will allow for owners of latent computing resources to be compensated in cryptographic tokens, which will strongly facilitate the participation of computing resources in the network, while still keeping them cheap and affordable for users. Furthermore,
NuNet will enable truly decentralized cooperation and competition of AI models developed and trained by different researchers, individuals, and possibly economic actors (initially via SingularityNET AI network and ecosystem). For example, participants will wrap their trained models into SingularityNET services and offer them in the marketplace. NuNet will enable these models to be run on latent computing resources supplied to the network and compensate each supplier according to their contribution. Furthermore, NuNet will allow data needed for training and updating models to be easily accessed and compensated.
Mobile IoT device ecosystems and smart-city implementations
Mobile IoT device ecosystems, such as sensors and cameras equipped drones, cars, smartphones, and in general more or less advanced autonomous robots provide implementation challenges simply due to the fact that their topologies constantly change. Furthermore, network connectivity speeds and patterns may change considerably when components of the network move with respect to each other. NuNet, leveraged by the AI ecosystem of SingularityNET, provides the ability to balance computing loads between the edge and ‘core’ of such networks and subnetworks thus supporting diverse mobile or stationary IoT device ecosystems, such as semi-autonomous rescue and security drone fleets, car fleets, collaborative robots, truck platoons, etc. Furthermore, NuNet enables cross-vendor cooperation via its tokenomics mechanism, allowing to the integration of devices and ecosystems of different vendors into a single computing workflow.
Cross-vendor process integration
Decentralized by design computing architectures and data workflows of IoT networks, which span large geographical areas and involve diverse ecosystems of individually secured devices, allows solution providers to integrate devices and computational processes owned and operated by different businesses into a single business process. Using blockchain-based custom state-of-the-art data privacy, provenance, access management solutions, and an economic mechanism powered by fine-grained microtransactions, NuNet and SingularityNET enable data economy and business ecosystems with many partners that do not need to be centrally managed or rely on a single trusted party. The capability of integrating multiple vendors and businesses into one value chain has huge potential in largely untapped IoT domains such as smart cities, international supply chains, and the management of large partnerships in general.