Internet Of AI Technology
Core concepts validated by scientific community in computer science journals
Conference paper - 08.07.2022
Artificial Intelligence Trends in Systems
Performance Analysis of an Ontology Model Enabling Interoperability of Artificial Intelligence Agents
Abstract - In the context of current information systems, the necessity of processing vast amounts of information enforces a significant paradigm shift. The monolithic data formats are gradually replaced with semantic models enclosing light, layered, and easily replicable structures. Advances represented by blockchain, decentralized data storage and replication along with semantic models open new possibilities to combine Artificial Intelligence (AI) algorithms in multi-agent, cross-organizational systems. By taking a semantic model approach, the communication and coordination between agents can be abstracted by the ontology concepts used to define the data models. This paper explores the feasibility of utilizing ontology-based semantic data models as communication interfaces among multiple AI agents. Specifically, it focuses on the storage, retrieval, and replication part of this process in a decentralized medium.
Conference paper - 16.07.2021
Artificial Intelligence in Intelligent Systems
A Decentralized Ontology Versioning Model Designed for Inter-operability and Multi-organizational Data Exchange
Abstract - Current information systems operated in a multi-organization context use centralized or hierarchical data models proven inefficient as the network’s scale rapidly increases. The introduction of blockchain and decentralized specific mechanisms stimulate the use of ontology-based data models. While the adoption of blockchain as a communication medium facilitates the communication and data exchange in a more reliably and securely fashion, this opens new challenges regarding the evolution and versioning of the ontology. This paper proposes a versioning mechanism tailored to the decentralized environment. It provides the conceptual building blocks allowing an ontology to evolve under the supervision of a community of individuals and organizations proposing progressive refinements to the ontology elements. Also, it details the aggregation mechanism used to select the changes for the new version.
Conference paper - 09.08.2020
Artificial Intelligence and Bioinspired Computational Methods
An Ontology Model for Interoperability and Multi-organization Data Exchange
Abstract - Progress in the uptake and use of technologies such as Artificial Intelligence and the Internet of Things seems to be stalled by the colossal fragmentation of information and data standards. This complexity is compounded by issues of inter-organisational differences, hindering effective collaboration. There is a growing demand for cross-organizational integrations in regulated but decentralized environments. This paper introduces an ontology architecture where information is sliced into independent semantic layers, each focusing on a specific aspect of the data. By dissolving traditional monolithic data structures into layered, light semantic components, the necessity to maintain contextual metadata is diminished. The model proposes an ontology structure based on decentralized technologies to ensure an open environment and to avoid the flaws of centralized systems.
Conference paper - 08.08.2020
Applied Informatics and Cybernetics in Intelligent Systems
Dynamic Consensus: Increasing Blockchain Adaptability to Enterprise Applications
Abstract - Decentralization powered by blockchain is validated for its capability to build trust like no other computational system before. The evolution of blockchain models has opened new use-cases that are becoming operational in many industry fields such as: energy, healthcare, banking, cross-border trade, aerospace, supply chain, and others. The core component of a decentralized architecture is the consensus algorithm - the set of rules that ensures an automated and fair agreement between the actors in the same network. Classic consensus algorithms are tailored to solve specific problems, but in an open ecosystem, each business case is unique and needs a certain level of customization. This paper introduces a new meta-consensus model called Dynamic Consensus, an architecture extension that allows multiple, complementary, consensus algorithms to run on the same platform. While classic consensus mechanisms are more appropriate for public or private systems (narrow set of rules), a dynamic approach would fit better for federated business consortiums (more rules and higher need for adaptability). The model is illustrated and analyzed as an ongoing experimental feature that can be added to enterprise blockchains designed to operate in cross-domain environments.
Conference paper - 08.05.2019
Software Engineering Methods in Intelligent Algorithms
The Evolution of Blockchain Virtual Machine Architecture Towards an Enterprise Usage Perspective
Abstract - Virtualization in the context of blockchain systems represents an essential phase in the development and migration of services from public chains to enterprise logic. Most of the ongoing blockchain uses-cases are using the existing public ledgers, but for business products and services, there is a need for custom tailored solutions to ensure flexibility and security. The Ethereum Virtual Machine has opened new ways to solve problems that require a public proof by executing logic on a decentralized ecosystem. In a natural evolutive process, virtualization logic was shaped by numerous architectures and business requirements. Beside performance and scalability, enterprise virtual machines are starting to focus on features like: formal verification, execution of business logic, cross-domain rules, upgradable smart contracts, data privacy politics and audit tools. The current paper presents an introduction, an architectural walkthrough of the most relevant virtual machines that currently exist in the enterprise blockchain ecosystem and an analytical perspective on the future of industry ready VM’s.
Technology Overview
Economy of
Innovation
In order to achieve high-quality, valuable, and reliable results, the support of an economic environment is required to cover innovators’ expenses through the monetization of their work. By satisfying the financial aspect, innovators can then dedicate their time and effort toward exploring, formulating, and creating elaborate solutions which accelerate the ecosystem's growth.
Service
Quality
In Openfabric, a Bayesian reputation model supervises the quality and performance of products and services through a reputation score that is computed based on community feedback. It also serves to increase collaboration amongst participants in a safe environment, without relying on a centralized authority.
Decentralized
Marketplace
The Openfabric marketplace provides a uniform, intuitive, and simplified user experience, which allows the execution of AIs without the necessity of installing, configuring, or customizing anything.
It consolidates the business relationship between the supply-and-demand of AI services, innovators, infrastructure providers, end-users, and businesses.
Privacy
For All
Privacy is an essential attribute of Openfabric, which stems from the fact that algorithms and data sets are decrypted only inside the TEE so that neither the platform nor the executor has access to it.
Enterprise
Integration
Considering enterprise adoption of edge technologies is slow, expensive and disruptive, Openfabric provisions connectors minimising the integration friction.
Decentralization
The distributed ledger ensures undeniable contracts and unforgettable history between the platform's stakeholders. It also serves as the underlying layer for access control and identification mechanisms. The platform is orchestrated by a decentralized operating system (DOS) which manages network resources, services, and processes to coordinate the proper functioning of the system.
Equilibrium
Nash equilibrium is achieved when infrastructure providers offer excellent services, innovators generate high-quality algorithms the community is willing to pay for, and service consumers efficiently combine algorithms to obtain solutions for their specific use cases.
Economy of Innovation
In order to achieve high-quality, valuable, and reliable results, the support of an economic environment is required to cover innovators’ expenses through the monetization of their work. By satisfying the financial aspect, innovators can then dedicate their time and effort toward exploring, formulating, and creating elaborate solutions which accelerate the ecosystem's growth.
Service Quality
In Openfabric, a Bayesian reputation model supervises the quality and performance of products and services through a reputation score that is computed based on community feedback. It also serves to increase collaboration amongst participants in a safe environment, without relying on a centralized authority.
Decentralized Marketplace
The Openfabric marketplace provides a uniform, intuitive and, simplified user experience, which allows the execution of AIs without the necessity of installing, configuring, or customizing anything.
It consolidates the business relationship between the supply-and-demand of AI services, innovators, infrastructure providers, end-users, and businesses.
Privacy For All
Privacy is an essential attribute of Openfabric, which stems from the fact that algorithms and data sets are decrypted only inside the TEE, so neither the platform nor the executor has access to it.
Enterprise Integration
Considering enterprise adoption of edge technologies is slow, expensive, and disruptive, Openfabric provisions connectors minimising the integration friction.
Decentralization
The distributed ledger ensures undeniable contracts and unforgeable history between the platform's stakeholders. It also serves as the underlying layer for access control and identification mechanisms. The platform is orchestrated by a decentralized operating system (DOS) which manages network resources, services, and processes to coordinate the proper functioning of the system.
Equilibrium
The distributed ledger ensures undeniable contracts and unforgeable history between the platform's stakeholders. It also serves as the underlying layer for access control and identification mechanisms. The platform is orchestrated by a decentralized operating system (DOS) which manages network resources, services, and processes to coordinate the proper functioning of the system.
Openfabric Architecture
Platform
Network
Openfabric Worker 1
Openfabric Worker N
Openfabric Worker 1
Openfabric Worker N
Openfabric Token
The Openfabric token is a utility token fueling the ecosystem economy by enabling value transfer between parties.
Staking
Reward
Payment
Medium of
Exchange
Governance
Challenging the Most Difficult
Problems Faced by AI Platforms
Decentralization
Ensure there is no central entity controling the location of data or information processing.
Usability
Simplify interactions between end-users and AIs by providing straightforward, nontechnical flows.
Security
Protect end-user privacy and guarantee intellectual property rights.
Smart Economy
Create a built-in robust exchange medium that facilitates fair transactions between supply-and-demand of AI services.
Interoperability
Implement the use of standardized interfaces, to allow multiple AI agents to cooperate and connect in order to provide relevant answers to complex problems.
Scalability
Expand network capabilities by allowing network participants to rent their computing power for the execution and training of AIs.
Decentralization
Ensure there is no central entity controling the location of data or information processing.
Usability
Simplify interactions between end-users and AIs by providing straightforward, nontechnical flows.
Security
Protect end-user privacy and guarantee intellectual property rights.
Smart Economy
Create a built-in robust exchange medium that facilitates fair transactions between supply-and-demand of AI services.
Interoperability
Implement the use of standardized interfaces, to allow multiple AI agents to cooperate and connect in order to provide relevant answers to complex problems.
Scalability
Expand network capabilities by allowing network participants to rent their computing power for the execution and training of AIs.
Openfabric Roadmap
Positioning
Brief comparison with the most notable alternatives suggests that Openfabric offers the most attractive set of features.
| Features | Centralized | Decentralized | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| IBM Watson | Google AI | Azure AI | Amazon ML | Singularity | Ocean Protocol | Effect AI | DeepBrain | Thought Network | Openfabric | ||
 | Governance |  | | | |  |  |  | |  | |
| Execution | | | | | | | | | | | |
| Storage | | | | | | | | | | | |
 | Ownership | | | | | | | | | | |
| Privacy | | | | | | | | | | | |
 | Scalability | | | | | | | | | | |
| Trustfulness | | | | | | | | | | | |
 | Developer-friendly | | | | | | | | | | |
| User-friendly | | | | | | | | | | | |
| Tools-integration | | | | | | | | | | | |
 | Fair market | | | | | | | | | | |
| Open market | | | | | | | | | | | |
| Marketplace | | | | | | | | | | | |
 | Algorithm composition | | | | | | | | | | |
| Structural level | | | | | | | | | | | |
| Open source Framework support | | | | | | | | | | | |
Decentralization
Security
Computation
Accessibility
Smart Economy
Interoperability
Decentralization
Security
Computation
Accessibility
Smart Economy
Interoperability
Openfabric integrations
Openfabric brings AI to Web3 by supporting all major blockchain ecosystems.
