INTEND - Intent-based data operation in the computing continuum 

ID Call: HORIZON-CL4-2023-DATA-01 Cognitive Computing Continuum: Intelligence and automation for more efficient data processing 

Sapienza's role in the project: Other beneficiary

Scientific responsible for Sapienza: Donatella Firmani

Department: Statistical Sciences

Project start date: January 1, 2024

Project end date: December 31, 2027

Abstract del progetto:

The EU's strategic autonomy in the digital economy requires more and more data to be processed in the Cloud-Edge-IoT computing continuum, instead of only the central cloud. This requires advanced automation and intelligence of the continuum. At the same time, recent breakthroughs in AI research have shown unprecedented intelligence to handle creative tasks. Such human-like intelligence will eventually disrupt how people use the cloud and continuum. INTEND aims at bringing such human-like intelligence into the cognitive continuum, to achieve the novel concept of intent-based data operation, which is capable of (1) continually learning how to manage and adapt heterogeneous cloud/edge resources for more efficient data processing, (2) strategic decision making across the decentralized continuum for end-to-end data security and sustainability, and (3) human-friendly interaction with data stakeholders in natural language, for effective and trustworthy human-entric data operation.

To achieve this ambitious objective, INTEND builds a strong consortium of world-class AI researchers and leading companies that cover the complete supply chain of the computing continuum. The project will deliver 11 novel software tools, which integrate into an INTEND toolbox. The approach and the tools will be tested and validated on 5 vertical domains, to achieve the novel intent-based data operation for video streaming pipelines, machine data platforms, 5G data infrastructure, urban data space, and robotic AI applications. An open platform will be provided to support the extension with new hardware, management tools and AI models into the cognitive continuum. The outputs pave the way for migrating the EU's data industry from the cloud to the continuum and implementing the EC's strategy of human-centric AI in the domain of data processing and computing continuum.

The contribution of Sapienza University will focus on the development of technologies for Knowledge Graphs (KGs) and generative Artificial Intelligence approaches. In particular, Sapienza will play a key role in building a comprehensive KG to represent the semantics of data operations and the intent of various stakeholders. Sapienza's role encompasses the following levels: (1) ensuring the extraction of relevant concepts from the KG, (2) addressing the issue of KG storage in the cloud/edge/IoT continuum, and (3) implementing privacy protection measures. Additionally, Sapienza will lead the development of methods based on large Language Models and generative Artificial Intelligence to extract human intents from DataOps artefacts. These methods will be able to generate human-readable documents to make the process of automatic updates more interpretable and facilitate collaboration between human stakeholders and automated agents for data processing.