We introduce a new abstract domain for analyzing memory block manipulations, focusing on programs with dynamically allocated arrays. This domain computes properties universally quantified over the value of the loop counters, both for assignments and tests. These properties consist of equalities and comparison predicates involving abstract expressions, represented as affine…
This talk will present MARVeLus, a cyber-physical systems language which combines verification, simulation, and implementation. The language is based on Zélus, a modern version of Lustre, where we add refinement types and associated typing rules. Our refinement types can express temporal-logic-based safety properties and prove them using a rich type…
The essence of abstract interpretation lies in the use of a simplified domain of abstract properties, known as the abstract domain, which enables the proof of properties about the concrete semantics of programs in a computationally feasible manner. In this seminar, I will demonstrate that certain elements within the concrete…
Temporal logics for hyperproperties have recently emerged as an expressive specification technique for relational properties of reactive systems. While the model checking problem for such logics has been widely studied, there is a scarcity of deductive proof systems for temporal hyperproperties. In particular, hyperproperties with an alternation of universal and…
In this talk, I will try to convince you that type checking can be seen as logic programs with a fine-grained input-output system, and query evaluations are proof reductions in classical linear logic. This gives us implementations of type checking algorithms for free, via direct translation from bidirectional typing rules…
In order to discover abstract interpretation, I implemented a well-known numerical abstract domain in the static analyzer MemCAD. This domain, described in 1974 by Karr, can infer affine equalities of the shape “aX + bY + … = c”. Such properties are helpful to develop optimizing compilers and static analyzers.…
I will describe work in progress on Osiris, an attempt to equip a nontrivial fragment of OCaml with a formal semantics and with an interactive program verification environment, hosted inside the Rocq proof assistant. Our semantics is hybrid: it takes the form of a monadic interpreter, defined on top of…
C is widely used in safety-critical systems, therefore it is important to have static analysis tools that can automatically identify program properties and thus help in the search for bugs. I will present C-2PO, a weakly-relational pointer analysis that can infer relationships between pairs of pointer terms in C programs,…
The causality in synchronous languages is a very old problem, used to produce modular scheduled imperative code, for which many answers have already been given. Interestingly, starting from a state-based semantics for a Lustre-like language, I will present how the causality could be used to ease the static analysis of…
Nous encodons dans Coq un modèle dénotationnel du langage d’entrée de Vélus, compilateur synchrone vérifié. Nous spécifions formellement les conditions de son équivalence avec le modèle relationnel existant, et nous explorons la possibilité d’un raisonnement interactif sur les programmes, à l’aide notamment d’un plongement du nouveau modèle dans celui des…
Many programs (e.g. malware) hide their behavior by using obfuscations such as opaque predicates. Automatic methods have been developed to detect such obfuscations. In this presentation, we will focus on static symbolic backward bounded execution, a method that enumerates backward bounded paths from a potential opaque predicate and uses symbolic…
Software analyzers and compilers have a lot in common: they both have to read and understand source code in order to prove facts about it and transform it into an equivalent code. While the goal of an analyzer is to prove facts (correctness, safety…) about the source, it often transforms…
Consensus, also known as Byzantine Agreement, has been extensively studied across various models, each characterized by different assumptions regarding cryptographic setups (such as private channels, public key infrastructure, hash functions, or common random strings), network synchrony (synchronous, partially synchronous, fully asynchronous), adversary capabilities (degree of adaptiveness, rushing behavior, computational limits),…
I will present the general notion of (weighted) averaging algorithms and the particular Metropolis algorithm which achieves consensus on the average of initial values in the case of networks with bidirectional communication links. This algorithm converges in quadratic time, even in the case of a dynamic network but, unfortunately, requires…
Due to its interdisciplinary nature, the development of data science code is subject to a wide range of potential mistakes that can easily compromise the final results. Several tools have been proposed that can help the data scientist in identifying the most common, low level programming issues. We discuss the…
In this talk, I will talk about how metaprogramming techniques can help build correct, flexible, and performant program analyzers. Metaprogramming techniques manipulate and transform programs as data objects. By considering static program analyzers as the objects to be manipulated or transformed, I will demonstrate how these techniques improve the construction…
Informal systems is one of the stewards of the open-source software of the Cosmos blockchain ecosystem. For instance, we are stewarding CometBFT, a state machine replication engine around Tendermint consensus. In addition, Informal is collaborating with other companies on the design and implementation of distributed systems and algorithms. In this…
Sound static analysis allows one to overapproximate all possible program executions to infer various properties. However, it requires quite some effort to formalize and prove the soundness of program semantics. Most software applications developed nowadays are distributed systems in which different computational entities communicate through synchronous and asynchronous mechanisms. These…
Developments in hardware have delivered formidable computing power. Yet, the increased hardware complexity has made it a real challenge to develop software that exploits hardware to its full potential. Numerous approaches have been explored to help programmers turn naive code into high-performance code, finely tuned for the targeted hardware. However,…
Academic research in static analysis produces software implementations. These implementations are time-consuming to develop and to maintain, in order to enable building further research upon the implementation. While necessary, these processes can be quickly time-consuming. This work documents the tools and techniques we have come up with to simplify the…
The aim of this thesis is the verification of task schedulers for operating systems through static analysis based on abstract interpretation. Operating systems are collections of software present on almost every computer. Their purpose is to allow other programs to run without having to manage low-level problems such as memory.…
Embedded control software is often specified as a set of periodic tasks together with a dependency graph describing their communications via shared variables. This talk will describe an ongoing collaboration with Airbus to express such systems in a specialized version of Lustre. This version of Lustre is “rate synchronous” in…
Dans cette présentation, je vous parlerai de mes travaux sur l’analyse de valeurs par interprétation abstraite des langages fonctionnels.Nos premiers résultats, basés sur des domaines abstraits relationnels et des résumés des champs récursifs des types algébriques, permettent d’analyser des fonctions récursives du premier ordre manipulant des types algébriques récursifs et…
Garbage collection greatly simplifies the life of the programmer, removing the need for manually deallocating memory. However, heap space usage of garbage-collected programs can be tricky to understand: to argue that the garbage collector can reclaim the space associated with a heap object, one has to prove that this object…
Digital filters are widely used in control/command programs: a digital filter is a numerical algorithm to smooth variations of sampled values of signal. Digital filters involve linear recursions that can hardly be bound by using linear invariants. We propose an abstract domain using quadratic constraints and formal expansion to bound…
We consider a certain dynamic system, that is, a system with a certain internal state whose evolution is ruled by a known state update function. At regular interval, a measurement is performed. Based on this series of measurements, our objective is obtaining a good estimation of the current state of…
Static analyses aim at inferring semantic properties of programs. We distinguish two important classes of static analyses: state analyses and relational analyses. While state analyses aim at computing an over-approximation of reachable states of programs, relational analyses aim at computing functional properties over the input-output states of programs. Relational analyses…
Modern programming languages often provide functions to manipulate regular expressions in standard libraries.If they offer support for advanced features, the matching algorithm has an exponential worst-case time complexity: for some so-called vulnerable regular expressions, an attacker can craft ad hoc strings to force the matcher to exhibit an exponential behaviour…
In the second part of the talk, we tackle the setting of dynamic networks and present a quite different approach for the distributed computation of functions, which is based on popular algorithms in statistical physics, namely the Metropolis algorithm and the Push-Sum algorithm. When an upper bound on the network…
In this talk, we present several results on distributed function computation in anonymous network with broadcast communications. First, we recall the characterization, given by Boldi and Vigna, of the computable functions when agents have no information on their outgoing neighborhoods. Then we characterize the class of computable functions in networks…
When studying categories in homotopy type theory, it turns out that the well-behaved definition is that of univalent categories, that is, categories such that isomorphisms between objects are equivalent to identities between them. It has been shown that any category is weakly equivalent to a univalent one; this operation is…
Violations of program properties pose significant risks, particularly when they can be triggered by attackers. Multiple approaches has been proposed recently to detect these vulnerabilities. In this seminar, we will present a new abstract interpretation-based static analysis for this problem. By leveraging a static analysis by abstract interpretation for CTL…
For performance and scalability reasons, most cloud providers start using Field-Programmable Gate Arrays (FPGAs) to provide FPGA-based acceleration services. FPGA virtualization becomes then necessary to optimize hardware resources utilization, but raises security and privacy concerns.In this work we consider a temporal multi-tenancy environment, where a single FPGA instance running a…
In this presentation, we are interested in the Computationally Complete Symbolic Attack (CCSA) approach to security protocol verification, and its implementation in the proof assistant SQUIRREL. The proof of a protocol relies on assumptions on the security provided by its primitives (e.g. hash, encryption, etc.), which are generally formulated in…
Conversion to Static Single Assignment (SSA) form is usually viewed as a syntactic transformation algorithm that gives unique names to program variables, and reconciles these names using “phi” functions based on a notion of domination. We instead propose a semantic approach, where SSA translation is performed using a simple dataflow…
I will present the work I did during my M1 internship in Dresden on two-variable logics (especially C², the two variable first-order logic with counting quantifiers), trying to show that every satisfiable sentence admits a model of finite cliquewidth.
One of the most important results in distributed computing was published in April 1985 by Fischer, Lynch and Paterson. Their short paper “Impossibility of Distributed Consensus with One Faulty Process” definitively placed an upper bound on what it is possible to achieve with distributed processes in an asynchronous environment. The problem of…
Probabilistic programming is a method for Bayesian statistics in which statistical models are encoded as programs. In this talk I will introduce probabilistic programming and the problem of inference. I will then present two kinds of semantics for probabilistic programs: one based on measures and probability kernels, and one based…
It is widely known that the precision of a program analyzer is closely related to intensional program properties, namely, properties concerning how the program is written. This explains, for instance, the interest in code obfuscation techniques, namely, tools explicitly designed to degrade the results of program analysis by operating syntactic…
We present an abstract interpretation framework for proving liveness properties of imperative programs, to ensure that some event happens once or infinitely many times during program execution. A further abstraction based on piecewise-defined ranking functions yields a static analysis that automatically infers sufficient conditions for these liveness properties as well as upper…
Fundamental concerns exist on the trustworthiness of deep learning systems, with examples including robustness, fairness, privacy and explainability. Phenomena like adversarial examples prompt the need to train robust networks and to provide formal guarantees on their behaviour. In this talk, we will first introduce the neural network verification problem. Then,…
We present the article “Shape analysis by means of bi-abduction” by Cristiano Calcagno, Dino Distefano, Peter O’Hearn, Hongseok Yang published at POPL in 2009. This article addresses the problem of pre- & post-condition generation and verification for function analysis. In order to resolve this problem, the authors introduce the notion…
Modeling paradigms for Systems Biology plays an important role in investigating the orchestrated function of various biological systems. Additionally, they permit one to study a system in-silico in order to gain mechanistic insights from the trajectories resulting from the model. A common battle to derive an ideal representation for a…
We present an approach to language design aimed at enforcing modular security, where a security architect can define specific object capabilities that enforce custom security properties. Unlike traditional models that assume an uncompromised system (‘clean garden’), this model (‘dark forest’) assumes local machines are already compromised but anchored by a…
Runtime errors that can be triggered by an attacker are sensibly more dangerous than others, as they not only result in program failure, but can also be exploited and lead to security breaches such as Denial-of-Service attacks or remote code execution. Proving the absence of exploitable runtime errors is challenging,…
In this talk, we will present an overview of the abstract domains that drive the Flambda 2 optimizer for In this talk, we will present an overview of the abstract domains that drive the Flambda 2 optimizer for OCaml programs. Like most optimizing compilers, Flambda 2 relies on static analysis…
Probabilistic programming languages aim at designing and implementing algorithms that compute over probability distributions rather than states. They feature dedicated primitives for sampling and conditioning. Moreover, the evaluation of probabilistic programs relies on general purpose inference engines. In the last few years, several languages such as Pyro, Edward or Stan…
I present two impossibility proofs for consensus in the presence of byzantine failures based on graph coverings.
Event structures, trace automata, and Petri nets are fundamental models in concurrency theory. There exist nice interpretations of these structures as combinatorial and geometric objects and both conjectures can be reformulated in this framework. Namely, from a graph theoretical point of view, the domains of prime event structures correspond exactly…
We tackle the problem of testing and verifying “incomplete” programs—that contain components whose source code is either unavailable (like third-party libraries and cloud-based APIs), or too complex to model formally (like large machine learning models). We show how classical techniques like symbolic execution, deductive verification and SMT solving can be…
