4th Workshop on
Artificial Intelligence and
Automata, and sYnthesis
The increasing adoption of Artificial Intelligence techniques in safety-critical systems, employed in real world scenarios, requires the design of reliable, robust, and verifiable methodologies. Artificial Intelligence systems employed in such applications need to provide formal guarantees about their safety, increasing the need for a close interaction between the Artificial Intelligence and Formal Methods scientific communities. To witness this increasing need, tools and methodologies integrating Formal Methods and Artificial Intelligence solutions are getting more and more attention.
The workshop is the main official initiative supported by the OVERLAY group. The event aims at establishing a stable, long-term scientific forum on relevant topics connected to the relationships between Artificial Intelligence and Formal Methods, by providing a stimulating environment where researchers can discuss about opportunities and challenges at the border of the two areas.
Important goals of the workshop are (i) to encourage the ongoing interaction between the formal methods and artificial intelligence communities, (ii) to identify innovative tools and methodologies, and (iii) to elicit a discussion on open issues and new challenges.
This year edition will be held on November 28th, 2022, during AIxIA 2022, which is scheduled to be held in Udine, Italy (unless it will be an online event, depending on the pandemic emergency situation). Precise dates will be announced as soon as possible.
Proceedings now available online
The workshop proceedings are now available online on CEUR-WS!
|Paper submission||October 14th, 2022 September 30th, 2022|
|Notification||October 28th, 2022 October 21st, 2022|
|Camera-ready||November 11th, 2022 November 4th, 2022|
|Workshop||November 28th, 2022|
University of Trieste, Italy
stl2vec: semantic preserving embeddings of signal temporal logic
Signal temporal logic is a powerful formalism to reason about real-time systems. Recently, a lot of work has been channeled towards requirement mining, i.e. the goal of learning STL specifications from data or from models. All these approaches, at least those not restricting to pre-specified templates, are substantially based on operations that manipulate the syntax of a formula. Yet it’s the semantics of the logic that carries the information that is needed. At the same time, the recent explosion of machine learning and deep learning showed that working with real vectors is much easier, and brought several strategies to learn encoding of combinatorial structures into a real feature space. In this talk, we will explore this direction for STL, taking however an approach that avoids to introduce uncontrollable approximations due to the use of neural networks, instead proposing an embedding based on suitable kernel functions. We will show how to construct implicit and explicit embeddings with a controllable quality (in a probabilistic sense) and test their quality on the problem of using machine learning to predict the result of model checking in a stochastic setting (i.e. to predict satisfaction probability of properties).
Call for contributions
We accept extended abstracts (4 pages + references) focusing on the interaction between Artificial Intelligence and Formal Methods. Invited talks will complement the presentations of contributed papers.
Topics of interest include (but are not limited to):
- automata theory
- automated reasoning
- automated planning and scheduling
- controller synthesis
- formal specification languages
- formal verification
- game theory
- hybrid and discrete systems
- logics in computer science
- reactive synthesis
- runtime verification and monitoring
- specification and verification of machine learning systems
- tools and applications
Contributed papers can present recent results at the border of the two fields, new research directions, challenges and perspectives. Presentation of results recently published in other scientific journals or conferences is also welcome.
We plan to include all papers in the Proceedings of the event, published at CEUR Workshop Proceedings . CEUR WS proceedings are archival proceedings indexed by DBLP and Scopus.
Submitted papers should not exceed four (4) pages plus references. Authors are asked to use the CEURART LaTeX style, available here.
Submissions must be in PDF format and will be handled via the EasyChair Conference system at the following address: https://easychair.org/conferences/?conf=overlay2022 .
Registration and participation fees
OVERLAY 2022 will take place as part of AIxIA 2022. As a consequence, participants to the workshop must be registered to AIxIA. The registration to the main conference gives access to all the workshops including OVERLAY.
See the main conference web site for information on registration fees and cutoff dates.
Taking into account the emergency situation, the organizers of OVERLAY 2022 will closely monitor the evolution of the pandemic and will take all the necessary actions to guarantee a successful event.
In the case that an in-person event will not be possible, OVERLAY 2022 will be a hybrid or fully virtual conference. Updates on the organization mode will be published and advertised in due time.
- Luca Geatti • Free University of Bozen-Bolzano
- Guido Sciavicco • University of Ferrara
- Alessandro Umbrico • ISTC-CNR, Rome
- Riccardo De Benedictis • CNR-ISTC, Rome, Italy
- Davide Bresolin • University of Padua, Italy
- Marco Esposito • Sapienza University of Rome, Italy
- Francesco Fabiano • University of Parma, Italy
- Marco Gavanelli • University of Parma, Italy
- Alessandro Gianola • Free University of Bozen-Bolzano, Italy
- Nicola Gigante • Free University of Bozen-Bolzano, Italy
- Federico Mari • University of Rome Foro Italico, Italy
- Andrea Micheli • Fondazione Bruno Kessler, Trento, Italy
- Emilio Munoz-Velasco • University of Malaga, Spain
- Adriano Peron • University of Naples Federico II, Italy
- Raine Ronnholm • University of Tampere, Finland
- Enrico Scala • University of Brescia, Italy
- Mauro Vallati • University of Huddersfield, United Kingdom
- Martin Zimmermann • Aalborg University, Denmark
For more information email to email@example.com.