1. Automated Verification Daniel Kroening Department of Computer Science, University of Oxford

2. What is Automated Verification?
Semantics of systems
Programming languages, and broader
Outside of CS: e.g., DNA computation, robotics
Algorithmic analysis of systems
Key challenge is incredible complexity of these systems
A computer that reasons about computers!
Area is transitioning!
At the heart of theoretical computer science, e.g., Hoare, Pnueli, Clarke, Lamport
Now becoming part of systems and programming languages (POPL, PLDI, ISSTA, OOPSLA)

3. What we do in Oxford
Theory
Application

4. Main Subareas Quantitative Verification HW/SW Analysis
Marta Kwiatkowska
Alessandro Abate
HW/SW Analysis
Daniel Kroening
Tom Melham
Hongseok Yang
Verification for Security
Bill Roscoe
Cas Cremers
Foundations of Verification
Stefan Kiefer
Luke Ong
Joel Ouaknine
James Worrell

5. Quantitative verification and synthesis
Focus on quantitative aspects of software/systems
probability, time, energy consumption, etc
multi-objective properties
Develop algorithms and software tools
automated verification and strategy/controller synthesis
parameter synthesis, towards model synthesis
Wide variety of case studies
cardiac pacemaker and heart modelling
energy smartgrid
DNA computation and robotics
Projects
VERIWARE (ERC Advanced Grant)
Mobile Autonomy (EPSRC Programme Grant)
Tools PRISM, PRISM-games, HeartVerify

6. Grants
Compositional higher order model checking: logic, models and algorithms
EPSRC
Ong
630k
Consolidator grant
ERC
Ouaknine
1411k
Verification of linear dynamical systems
Worrell
1005k
Reducing Cost of Software: A Scalable Model-Based Verification Framework
Roscoe
961k
Security and Privacy in Smart Grid Systems: Countermeasure and Formal Verification
Martin
202k
VERIWARE: From software verification to everyware verification
Kwiatkowska
1648k
Validation of Concurrent Software Across Abstraction Layers
Kroening
1094k
Mobile Autonomy: Enabling a Pervasive Technology of the Future
944k
Quantitative Analysis of Infinite-State Systems
Royal Society
Kiefer
419k

7. Indicators of Esteem Wolfson Research Merit Award - Daniel Kroening
Roger Needham Award - Joel Ouaknine
Will host FLOC 2018 in Oxford

8. Major Collaborations Toyota Siemens Programming Research BTC-ES
Grant funding for a decade
In-house use of our tools
Siemens
Product with CBMC as engine
Programming Research
Product with CPROVER-based engine
BTC-ES
Rapita Systems, Texas Instruments, Intel, Tata Consultancy Services

9. Impact Impact through tools! Diffblue
PRISM: analyzer for probabilistic models
CBMC: analyzer for C code
Diffblue
aims at commercialising research in computer aided verification – and in particular, the CBMC model checking system for automatically checking C code

10. PRISM PRISM: Probabilistic symbolic model checker
developed at Birmingham/Oxford University, since 1999
free, open source software (GPL), runs on all major Oss
simple but flexible high-level modelling language
Various efficient model checking engines and techniques
symbolic methods (binary decision diagrams and extensions)
explicit-state methods (sparse matrices, etc.)
statistical model checking (simulation-based approximations)
Graphical user interface
editors, simulator, experiments, graph plotting
See:
downloads, tutorials, case studies, papers,

11. Plans and Strategy From Verification to Synthesis
probability, time, energy consumption, etc
multi-objective properties
advantage: correct-by-construction
strategy synthesis: “can we construct a strategy to guarantee that a given quantitative property is satisfied?”
instead of “does the model satisfy a given quantitative property?”
parameter synthesis: “find optimal value for parameter to satisfy quantitative objective”
Many more application domains
robotics, control, power management, etc

12. Cardiac pacemaker Develop model-based framework Properties
timed automata model for pacemaker software
hybrid heart models in Simulink, adopt synthetic ECG model (non-linear ODE)
Properties
(basic safety) maintain beats per minute
(advanced) detailed analysis energy usage, plotted against timing parameters of the pacemaker
parameter synthesis: find values for timing delays that optimise energy usage
Synthesising robust and optimal parameters for cardiac pacemakers using symbolic and evolutionary computation techniques. Kwiatkowska, Mereacre, Paoletti and Patane, HSB’16

13. DNA origami tiles
DNA origami tiles: molecular breadboard [Turberfield lab]
50nm
Aim to understand how to control the folding pathways
formulate an abstract Markov chain model
obtain model predictions using Gillespie simulation
perform a range of experiments, consistent with preditions
AFM atomic force microscopy, scale bar 50nm; roughly 2,500 nucleotides (=825nm at0.33nm per nt), staples length nucleotides
The tile is created by annealing a 2646-nucleotide (nt)
circular template with 90 short staple oligonucleotides. The form of the folded template is
determined by the staples, 76 of which bind to two non-contiguous 15- or 16-nt domains of the
template and thus force them together to form a rectangular tile.
50nm
Guiding the folding pathway of DNA origami. Dunne, Dannenberg, Ouldridge, Kwiatkowska, Turberfield & Bath, Nature (in press)

14. The challenge of mobile autonomy
Autonomous systems
are reactive, continuously interact with their environment
including other components or human users, adversarial
have goals/objectives
often quantitative, may conflict
take decisions based on current state and external events
must adapt, reconfigure, etc
Natural to take a game-theoretic view
Many occurrences in practice
e.g. security protocols, distributed consensus, energy management, sensor network co-ordination, semi- autonomous driving, interacting with robotic assistants, …