• BBCA

Is AI on Chain the future?

AI and Blockchains are two of the newest technologies that are conquering the web with their applications. However, is the mixture of these two technologies the final solution to most of our modern problems?

In late 2017 and early 2018, The SingularityNET Foundation was formed with the dream to create a new AI fabric that delivers superior practical AI functionality. SingularityNET provides an automated process that enables any business to connect existing AI tools to build the solution it needs. In the beginning, the foundation will populate the network with its own AI Agents, discerning between "core AI algorithmic services" and "high-level AI services". At the most abstract level, one can think of an AI service as a function with a set of inputs and outputs. The service could be a low-level service that does a specialized unit of work or a higher-level service that calls upon a series of lower-level services to complete its task. Many services will require powerful computing resources and will, therefore, work in the cloud. This way allows the network to be lightweight and perfectly doable also in low-power devices. On SingularityNET, the exchange currency is the AGI Token. With it, Token holders can purchase services in the marketplace and have voting rights in the network's democratic governance system.

For end-users who want to purchase access to the AI services available in the platform, the most relevant component is the Marketplace DApp, through which they can search and browse a collection of curated services for a large and ever-growing variety of AI tasks.

The SingularityNET reputation system will help Agents (whether human or themselves AIs) choose between the many AI services that perform the same function. Rating system design is complex since it is multidimensional then it will require to evolve with time. Some of the different dimensions are general service performance, timeliness, accuracy and value for money. But for a practical purpose, it is valuable to have a single-digit rating between 0 and 5. The rating system is critical for making choices about everyday transactions in the network, but it also plays a core role in network governance and resource allocation.

The SingularityNET platform currently depends on the Ethereum blockchain. To improve scalability and expand adoption, it will support other existing blockchain technologies. The Registry and the Multi-Party Escrow are the two critical Smart contracts. The Registry is where AI service providers register on the platform, providing text descriptions and tags to allow users to discover their service, pricing information and endpoint locations to let users call their services. The Multi-Party Escrow contract handles payments through escrow accounts for each user (end users and applications) for faster and cheaper transactions.

The SingularityNET's consensus algorithm is an evolution of the PoS (Proof-of-Stake) called Proof of Reputation (PoR), which combines several factors: stake, activity in the network, specific rating aspects and others. To optimize the combination of factors can be used Machine learning algorithms. In PoR, each participant maintains a distributed ledger of reputation evidence, which achieves consistent with the consensus algorithm. Compared with the Proof of Work (PoW), the consensus algorithm of Bitcoin, this protocol uses reputation as the incentive, which is cost-efficient since there are no miners or hash power during block competition. The cryptographical nature of blockchain can protect the integrity and reliability of the reputation evidence.

The first four areas of interest on high-level AI in the SingularityNET Platform are Network analysis, social robotics, bio-data analytics, probabilistic graphical models and serious games.

A Network analysis example would be a natural language tool to reduce the human labour of putting data into applications. This tool will interpret natural language texts (such as medical research papers) and summarize them to simplify the process of discovery. Researchers designed it to nudge unsupervised clusters into a human-designed ontology through seeding with a few exemplars using networked relations.

Instead, Social robotics will improve the well-being of humans through the use of natural interfaces and artificial intelligence. Instead of adapting human behaviour and society to technology, this adapts technology to meet natural human behaviour, creating social and cross-culturally intuitive interfaces.

One application of bio-data analytics could be the rapid identification of specific biomarkers that indicate alterations in the health condition. The discovery of new biomarkers provides direct support in observational and analytic epidemiology, clinical trials, screening, diagnosis, and prognosis.

Deep-reinforcement-learning methods have lately become some of the most popular algorithms in AI. However, outside of the gaming environment, they are not applicable. One possible use would be to find the best treatment for a patient with a complicated history or to work with practically any real-world data.

In the current economic and technological context, every business needs AI, but off-the-shelf AIs will rarely match the requests. Only tech giants can hire armies of developers to build custom AIs, and even they have a hard time hiring enough AI experts to meet demand. In a world like this, platforms like SingularityNET are essential for a business to grow faster and deliver a better user experience. At the same time, as we have seen, these new technologies can let humanity move forward in an era with better healthcare and a combination of intelligence, human and artificial one.

Author: Nicolò Stranieri

139 views0 comments

Recent Posts

See All