First Workshop on
Artificial Intelligence and
fOrmal VERification, Logic,
Automata, and sYnthesis

OVERLAY @ AIxIA 2019
November 19-20 2019, Rende (Italy)

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. AI systems employed in such applications need to provide formal guarantees about their safety, increasing the need for a synergic collaboration between the AI and Formal Methods scientific communities. Despite this increasing need, tools and methodologies integrating Formal Methods and Artificial Intelligence solutions have received relatively little attention.

The workshop is the first official initiative supported by OVERLAY, presenting the research group and its current results to the Italian AI scientific community. 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 FM and AI communities, (ii) to identify innovative tools and methodologies, and (iii) to elicit a discussion on open issues and new challenges.

Important dates

Please note the submission deadline has been extended to September 26th!
  • Paper submission: September 6th September 21st September 26th, 2019
  • Notification: September 28th October 9, 2019
  • Camera-ready: October 15th November 1st, 2019
  • Workshop: November 19-20, 2019

Invited speaker

Alessandro Abate
Oxford University
Title: Certified Reinforcement Learning with Logic Guidance

A model-free Reinforcement Learning (RL) framework is proposed, to synthesise policies for an unknown, and possibly continuous-state, Markov Decision Process (MDP), such that a given linear temporal property is satisfied.

We convert the given property into an automaton, namely a finite-state machine expressing the property. Exploiting the structure of the automaton, we shape an adaptive reward function on-the-fly, so that the RL algorithm can synthesise a policy resulting in traces that probabilistically satisfy the linear temporal property.

Under the assumption that the MDP has finite number of states, theoretical guarantees are provided on the convergence of the RL algorithm. Whenever the MDP has a continuous state space, we empirically show that our framework finds satisfying policies, if existing. Additionally, the proposed algorithm can handle time-varying periodic environments.

The performance of the proposed architecture is evaluated via a set of numerical examples and benchmarks, where we observe an improvement of one order of magnitude in the number of iterations required for the policy synthesis, compared to existing approaches (when available).

Speaker's bio

Alessandro Abate is Professor of Verification and Control in the Department of Computer Science at the University of Oxford (UK), and is a fellow of the Alan Turing Institute in London (UK). He received a Laurea in Electrical Engineering in October 2002 from the University of Padova (IT), an MS in May 2004 and a PhD in December 2007, both in Electrical Engineering and Computer Sciences, at UC Berkeley (USA). He has been an International Fellow in the CS Lab at SRI International in Menlo Park (USA), and a PostDoctoral Researcher at Stanford University (USA), in the Department of Aeronautics and Astronautics. From June 2009 to mid 2013 he has been an Assistant Professor at the Delft Centre for Systems and Control, TU Delft - Delft University of Technology (NL).

Call for contributions

We elicit the contribution of extended abstracts (4 pages + references) discussing the interaction of Artificial Intelligence and Formal Methods. Invited talks will complement the presentations of contributed papers.

Topics of interest include (but are not limited to):

  • automated reasoning
  • automated planning and scheduling
  • controller synthesis
  • formal verification
  • formal specification languages
  • game theory
  • hybrid and discrete systems
  • logics
  • reactive synthesis
  • runtime verification and monitoring
  • specification and verification of machine learning systems
  • timed automata
  • 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 welcome.

All papers will be included in the Proceedings of the event, published in the CEUR Workshop Proceedings AI*IA Series as well as on the workshop website. CEUR WS proceedings are archival proceedings indexed by DBLP and Scopus.

SUBMISSIONS

Submitted papers should not exceed four (4) pages, not including references. Authors are asked to use the workshop's LaTeX style.

Submissions must be in PDF format and will be handled via the EasyChair Conference system at the following address: https://easychair.org/conferences/?conf=overlay19

Program Committee

Chairs
PC Members (expanding...)

Support

We acknowledge the support of the Department of Mathematics of the University of Padova.

Department of Mathematics - University of Padova