Artificial Intelligence has been the common thread of my work for more than 25 years — from my Master thesis on reinforcement learning in 2000, through knowledge representation and reasoning for the Semantic Web, to learning from data streams.
In 2007 I started the Stream Reasoning research field, positioning it at the intersection between Stream Processing and Artificial Intelligence, and I have organized every edition of the Stream Reasoning Workshop series since. The extensions I proposed to the Semantic Web stack — RDF streams and Continuous SPARQL — were brought to the W3C RDF Stream Processing community group.
My work has been applied to space, manufacturing, IoT, mobile telecom and social media data, in collaboration with Leonardo, Thales Alenia Space, Siemens, Indra, Telecom Italia, Equinor and IBM.
Research lines
Five lines, each with an extended description and a selection of works with my comments, links and BibTeX.
Streaming Continual Learning (2022–)
Learning from non-stationary streams with temporal dependence: adapting fast to change without forgetting what came before. The meeting point of Streaming Machine Learning and Continual Learning.
Inductive Stream Reasoning (2008–)
What can be learned from a stream while it flows: Streaming Machine Learning, Time Series Analytics and Edge AI.
Deductive Stream Reasoning (2008–)
Where Stream Reasoning started: RDF streams, Continuous SPARQL and RSP-QL, Graph Stream Processing and Complex Event Processing.
Data Science and Engineering at Scale (2013–)
What survives contact with real volumes: distributed stream processing, complex event recognition, urban data science, and data and knowledge with Large Language Models.
Web Information Management (2001–2014)
The earliest line, now closed: Web and Information Retrieval, Web Data, Semantic Search, Semantic Web Services.
Books
- Streaming Artificial Intelligence — Springer, Data-Centric Systems and Applications, forthcoming 2027. With F. Giannini and R. Tommasini. It follows the structure of my Streaming Data Analytics course, whose courseware is open.
- Streaming Linked Data: From Vision to Practice — Springer, 2023. With R. Tommasini, P. Bonte and F. Spiga.
- Relevant Query Answering over Streaming and Distributed Data — Springer, 2020. With S. Zahmatkesh.
- Web Information Retrieval — Springer, Data-Centric Systems and Applications, 2013. With S. Ceri, A. Bozzon, M. Brambilla and P. Fraternali.
- Semantic Web: dai fondamenti alla realizzazione di un’applicazione — Pearson, 2009. In Italian.
Tools
- CapyMOA (2024–) — an open-source Python library for efficient Streaming Machine Learning, led by H. M. Gomes and A. Bifet. My group is among its most active contributors, extending it towards learning from streams with temporal dependence and Streaming Continual Learning. Together with Avalanche, it was one of the two hands-on frameworks of the 1st Streaming Continual Learning Bridge at AAAI 2026.
- MOA — Massive Online Analysis, the Java framework for mining data streams developed at the University of Waikato, and the engine CapyMOA builds on.
- Tenet (2024–) — the first publicly available benchmark for evaluating data stream classifiers in the presence of temporal dependence.
- RSP4J (2016–) — a second-generation library to build RDF Stream Processing engines following the RSP-QL reference model, and the official repository for the examples in the Streaming Linked Data book.
- C-SPARQL Engine (2008–2015) — the first RDF stream processor with stream reasoning capabilities.
The full record
These pages give context, not a complete list. For everything I have published, see DBLP, Google Scholar or ORCID.