abstract
The project aims to extend the capabilities of the JaKtA framework by introducing the capacity for the agent to have a Web-interactive environment. The primary goal is to provide agents with the ability to make HTTP calls to online resources, interpret responses in various formats, in our case (HTML, YAML, RDF), and specifically query the received data using specific query languages. The core of the project is the design and development of a library of concrete actions that each agent can register and execute. These actions include, for example, the ability to make a GET, POST, PUT, or DELETE request to a specific URL, execute CSS queries on an HTML page, extract structured values from a YAML document using JSONPath, or query an RDF graph with SPARQL. Each action is encapsulated in a modular structure that conforms to the framework’s abstractions, with inputs and outputs. The envisaged use cases involve scenarios in which an autonomous agent needs to extract information from web pages, APIs, or semantic datasets. For example, consider an agent querying a public RDF endpoint to learn about an author’s work, or an agent reading a YAML configuration from a remote resource to adapt its behavior. The project offers a significant contribution both in terms of semantic interoperability and in terms of cognitive agents capable of actively interacting with the Web. This extension can find applications in domains such as Web Intelligence, Semantic Web, and Intelligent Automation based on dynamic data.
outcomes
