CE23 - Intelligence artificielle

Hypermedia Communities of People and Autonomous Agents – HyperAgents

HyperAgents: Hypermedia Communities of People and Autonomous Agents

The HyperAgents project aims at enabling the deployment of world-wide hybrid communities of people and autonomous agents on the Web. For this purpose, HyperAgents defines a new class of multi-agent systems that use hypermedia as a general mechanism for uniform interaction such that they are (i) aligned with the Web architecture to inherit the properties of the Web as a world-wide, open, and long-lived system, (ii) transparent and accountable to support acceptance by people.

World-wide, open, long-lived, transparent and accountable hybrid communities of people and autonomous agents on the Web

To meet the needs of the hybrid communities, Hypermedia MAS need to address the following issues: <br />- world-wide: Hypermedia MAS have to accommodate billions of people and autonomous agents, to scale up to the size of the Web, and to cope with the high latency inherent in world-wide systems (e.g., real-time interactions among agents may no longer be feasible); <br />- open: Hypermedia MAS have to cope with the independent deployment of components (autonomous agents, tools, knowledge repositories, organizations, datasets etc.), and with the heterogeneity inherent in components that are developed independently from one another; <br />- Long-lived: Hypermedia MAS have to cope with the evolution of components over long periods of time, and to support backward- and forward-compatibility (i.e., old and new implementations have to coexist in one system); <br />- transparent and accountable: Hypermedia MAS thus have to the expected benefits of Hypermedia MAS may be negated by the effect of a variety of intentional, accidental or error risks raised by autonomous agents. Hypermedia MAS thus have to support transparency and accountability of agents by design <br />In order to enable the deployment of world-wide hybrid communities taking into account these issues, the following objectives are considered: <br />(i) Clear-cut design rationale: to define a set of design principles for Hypermedia MAS that induce the required architectural properties, <br />(ii) Interaction as first-class abstraction: to create declarative languages and hypermedia-based mechanisms for specifying and enacting interaction protocols for people and autonomous agents in Hypermedia MAS (and on the Web in general), so that they are able to publish, discover, interpret, and enact interaction protocols, and when applicable to modify the protocols as they evolve throughout their life-cycle; <br />(iii) Regulation as first-class abstraction: to create declarative languages and hypermedia-based mechanisms for specifying, monitoring, and enforcing regulations for people and autonomous agents in Hypermedia MAS (and on the Web in general) so that people and autonomous agents are able to discover and interpret regulations, and when applicable to publish, modify, monitor, and enforce the regulations as they evolve throughout their life-cycle; <br />(iv) Evaluation metrics and prototypical deployments: to define metrics for evaluating Hypermedia MAS and to demonstrate the applicability of the proposed contributions through prototypical deployments of hybrid communities.

Expertise on Distributed artificial intelligence, autonomous agents and MAS is central to address the defined objectives while overcoming the misalignment between Web-based MAS and the Web architecture on one hand and the definition of Large-scale, open MAS with humans in the loop on the other hand,
In conjunction with this expertise, expertise in Distributed hypermedia systems and the Web architecture is also essential in the achievement of the objectives while taking into account the challenging issue raised by the misalignment between Web-based MAS and the Web architecture.
Finally, Knowledge engineering, Semantic Web, and standardization are required expertise to define a unified set of vocabularies for representing Hypermedia MAS on the Web. HyperAgents will use and extend well-established knowledge representation formalisms and ontologies.
The work proposed to achieve objectives and to produce results will investigate the modelling and architecting of Hypermedia MAS, and of their interaction and regulation interconnected processes. They will be developed incrementally, ensuring the continuous integration and evaluation of results at the end of each phase.

The project aims to deliver results:
- Architectural style and reference architecture for Hypermedia MAS: a set of architectural design choices that constrain the roles and features of components and connectors within any architecture for Hypermedia MAS in order to ensure compliance to requirements; a reference architecture will also be provided
- Ontologies and declarative languages for Hypermedia MAS: a unified set of ontologies for representing Hypermedia MAS on the Web, together with a set of declarative languages for specifying interaction and regulation in Hypermedia MAS;
- Open-source software infrastructure for hybrid communities: open-source implementations of software components, libraries, and tooling for hybrid communities (as Hypermedia MAS); the infrastructure will conform to the proposed architectural style and reference architecture; the tooling will support researchers and practitioners to design, deploy, and manage hybrid communities on the Web;
- Prototypical deployments of hybrid communities: the proposed Hypermedia MAS will be used to deploy two prototypical hybrid communities for: (1) distributed industrial manufacturing, which will be deployed across two geographically distributed laboratory production cells -- one in France and one in Switzerland, and (2) tackling online disinformation, which will be deployed during the project and will remain online after its end as a demonstrator of the results of the project;
- Contribution to Standardization Efforts.

continue the project

updated all along the project life

The HyperAgents project aims to enable the deployment of world-wide hybrid communities of people and autonomous agents on the Web. For this purpose, HyperAgents defines a new class of multi-agent systems (MAS) that use hypermedia as a general mechanism for uniform interaction such that they are: (i) aligned with the Web architecture to inherit the properties of the Web as a world-wide, open, and long-lived system, and (ii) transparent and accountable to support acceptance by people. We refer to this new envisioned class of Web-based MAS as Hypermedia MAS. The HyperAgents project proposes: (1) to define an architectural style for Hypermedia MAS that induces the above-mentioned properties, (2) to define declarative languages and mechanisms for specifying, enacting, and regulating interactions among people and autonomous agents in Hypermedia MAS, (3) to develop an open-source software infrastructure for Hypermedia MAS that enables the deployment of hybrid communities on the Web, and (4) to demonstrate the deployment of prototypical hybrid communities in two application areas: (i) Industry 4.0 and (ii) tackling online disinformation. To undertake this investigation, the project consortium brings together internationally recognized researchers actively contributing to research on autonomous agents and MAS, the Web architecture, Semantic Web, and to the standardization of the Web. Being able to harness the collective power of Web-scale hybrid communities and to focus it on specific problems would have extensive implications in a broad range of domains: for manufacturers, it could enable distributed intelligent manufacturing at global scale; for citizens, it could enable access to transparent and trustworthy online information. We expect that the work conducted in this project will break new ground in AI and Web research and that its applications will cut across society.

Project coordinator

Monsieur Olivier Boissier (Laboratoire Hubert Curien)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

Laboratoire Hubert Curien
Inria Centre de Recherche Inria Sophia Antipolis - Méditerranée
HSG University of St. Gallen, Chair for Interaction- and Communication-based Systems

Help of the ANR 436,203 euros
Beginning and duration of the scientific project: February 2020 - 42 Months

Useful links

Explorez notre base de projets financés

 

 

ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter