asfenbits.blogg.se - Max 4 live otomata

#MAX 4 LIVE OTOMATA VERIFICATION#
#MAX 4 LIVE OTOMATA SOFTWARE#
#MAX 4 LIVE OTOMATA CODE#

TACAS, April 2016, Eindhoven, Netherlands

Synthesizing Piece-Wise Functions by Learning Classifiers.

An Automaton Learning Approach to Solving Safety Games over Infinite Graphs.

Invited Talk, May 2016, Universität Paderborn, Germany Invited talk, July 2016, Technische Universität Dortmund, Germany LDWA, September 2016, Hasso-Plattner-Institut, Potsdam, Germany

Synthesizing Invariants via Iterative Learning of Decision Trees.

LiVe (at ETAPS), April 2017, Uppsala, Sweden DARS (at CAV), July 2017, Heidelberg, Germany GI Theorietage, September 2017, Bonn, Germany

Solving Safety Games over Infinite Graphs.

Invited talk, February 2018, University of Kaiserslautern, Germany

#MAX 4 LIVE OTOMATA VERIFICATION#

Invariant Synthesis for Incomplete Verification Engines.Invited talk, June 2018, Saarland University, Germany Machine Learning Meets Formal Methods.Invited tutorial at MOVEP '18 Summer School, July 2018, ENS Paris-Saclay, France Invited talk, Highlights of Logic, Games and Automata, September 2018, Berlin, Germany Horn-ICE Learning for Synthesizing Invariants and Contracts.Invited talk, Complexity, Algorithms, Automata and Logic Meet, January 2019, Chennai Mathematical Institute, India Invited talk, March 2019, University of Bochum, Germany Formal Verification Meets Machine Learning.Invited tutorial at the ForMaL Spring School onįormal Methods and Machine Learning, June 2019, ENS Paris-Saclay, France

Deductive Verification, the Inductive Way.Sorcar: Property-Driven Algorithms for Learning Conjunctive Invariants.Invited talk, October 2019, University of California, Los Angeles, USA Combining Deductive and Inductive Reasoning in Formal Methods.Invited talk, January 2020, University of Liverpool, UK Learning Correctness Proofs of Software.Invited talk, November 2020, University of Oldenburg, Germany Logic and Learning: Formal Guarantees for Trustworthy Intelligent Systems.Invited talk, June 2021, RWTH Aachen University, Germany

#MAX 4 LIVE OTOMATA SOFTWARE#

(Horn-)ICE Learning: An Inductive Approach to Deductive Software Verification.

Invited talk, December 2021, Technical University of Dortmund, Germany

Safety and Explainability of Learning Systems.

Neuro-Symbolic Verification of Deep neural Networks.

To this end, the project combines inductive and deductive techniques from the areas of machine learning, artificial intelligence, and logic.Īs a byproduct, the project also investigates explainable machine learning, specifically learning of human-interpretable models. The goal of the DFG-funded project Temporal Logic Sketching is to develop computer-aided methods to assist engineers in writing formal specifications.

Horn-ICE: A learning-based program verifier.

flie: The Formal Language Inference Engine.

You can try some of the tools we have developed online: Our projects often extend into theoretical computer science, specifically to automata theory, game theory, and logic.

Learning theory: developing principled ways to combine learning and symbolic reasoning.

Incorporating automata learning and symbolic reasoning in reinforcement learning, inverse reinforcement learning, and transfer learning.

Specification learning/recommendation: learning-based methods that assist humans in writing formal specifications.

#MAX 4 LIVE OTOMATA CODE#

Learning-based synthesis: automatically generating reactive systems and program code from logical specifications and examples.Intelligent formal methods: tools that learn correctness and termination proofs for a wide range of systems, including hardware, software, and cyber-physical systems.Explanability of artificial intelligence: generating provably correct explanations for the decision making of intelligent systems.Verification of learning systems: automatically proving correctness properties (e.g., robustness) of learning systems.My group is working on a number of topics, including: The goal of my group is to build automated tools for the design, the construction, and the analysis of intelligent systems. I am especially interested in combining inductive techniques from the area of machine learning and symbolic techniques from the area of logic. My research interests broadly lie in the intersection of machine learning and formal methods. Moreover, I am principal investigator in several research projects. I am a research group leader at the Max Planck Institute for Software Systems in Kaiserslautern and teach at the University of Kaiserslautern (see below a list of course I offer). I have moved to the Carl von Ossietzky Universität Oldenburg, where I am now professor for Safety and Explainability of Learning Systems.