Agenda
février
-
13:0014:00
Vectors addition systems with states (VASS), a model equivalent to Petri nets, are finite-state machines with finitely many counters ranging over the natural numbers. The decidable reachability problem for VASS has many applications in logic, automata, and verification. With Jérôme and Grégoire, we studied the reachability problem for BVASS, a branching generalization of VASS. We show that BVASS reachability sets are very similar to VASS reachability sets, namely that they are sections of VASS. Our proof relies on a new well-quasi-order (wqo) on BVASS runs that generalizes the well-known wqo on VASS runs. By leveraging an amalgamation property, we prove that every BVASS run can be transformed into an equivalent one of bounded branching complexity. This allows us to derive several results on the geometry of BVASS reachability sets.
FrançaisLaBRI, salle 178 -
12:3013:00
Ce webinaire est animé par Naoufel Bahroun, Docteur en sciences de l’information, conservateur des bibliothèques, chargé de mission médiation et accompagnement à la Science ouverte – Service transversal de Soutien à la recherche (DIRDOC).
Il vise à identifier les pratiques typiques des éditeurs prédateurs et à sensibiliser les participants aux risques qu’ils représentent. Il propose des repères et outils concrets pour mieux les reconnaître, évaluer la fiabilité d’une revue ou d’un éditeur.Ce format court (30 min) a pour objectif de sensibiliser aux enjeux et pratiques de la science ouverte toute la communauté universitaire de Bordeaux, et particulièrement les chercheurs et doctorants.
-
14:3015:30
We introduce Sampled-Based Guided Quantum Walk (SamBa–GQW), a
novel quantum algorithm for approximating solutions of binary
combinatorial optimization problems of arbitrary degree with no use of
any classical optimizer. The algorithm is based on a continuous-time
quantum walk on the solution space represented as a graph. The walker
explores the solution space to find its way to vertices that minimize
the cost function of the optimization problem. The key novelty of our
algorithm is an offline classical sampling protocol that gives
information about the spectrum of the problem Hamiltonian. Then, the
extracted information is used to guide the walker to high quality
solutions via a quantum walk with a time-dependent hopping rate. We
investigate the performance of SamBa–GQW on several quadratic and
higher-order polynomial problems. We empirically demonstrate that
SamBa–GQW finds high quality approximate solutions on problems up to a
size of n=20 qubits by only sampling n^2 states among 2^n possible
decisions.FrançaisOn Zoom -
09:0010:00
Most of the quantum algorithms have been developed using a finite set of universal quantum gates, which allows for an easy decomposition of any unitary transformation. However, this set of gates includes entangling two-body gates. In the current context of faulty devices, the near-term noisy intermediate-scale quantum era, implementing these gates is deemed costly, both because the time requirement to implement these operations is much larger than for the single-qubit gates, and because they are more susceptible to errors. In contrast to the digital paradigm, analog quantum computers do not allow for arbitrary operations. But in turn of its lack of flexibility, analog devices excel in their robustness and resilience against noise.
In this context, digital-analog quantum computing (DAQC) was proposed as a way of taking advantage of the robustness of analog quantum computers while maintaining the flexibility of digital quantum computing. For this, this paradigm employs the natural interaction Hamiltonian of the devices as a resource for the entangling operations. Then, universality can be reached by the application of arbitrary single qubit gates. This allows us to implement any desired operation without the need of employing expensive and noisy two-qubit gates.
Here I present the contributions to the state of the art made during my PhD. In particular, I will describe the solution to one of the main challenges of DAQC, that was achieving universality employing efficient constructive methods. While optimal DAQC protocols require tackling an NP-Hard problem, we have proposed suboptimal protocols which only require polynomial classical resources to be calculated. In addition to this, we further studied its noise resilience, and introduced error mitigation techniques specifically designed for this paradigm. We further continued on the implementation of various algorithms and simulation tasks, showing the possibility that DAQC offers a valid alternative which in some cases might provide advantages in their run-time or fidelity.
FrançaisOn Zoom
mars
-
11:0012:00
Vincent Bridonneau (Université de Le Havre Normandie)
Title: Generation and analysis of dynamic graphs
Abstract:
This work investigates iterative processes that generate sequences of graphs, with applications spanning complex network analysis and temporal graph theory. Given a seed configuration, such processes yield a continuous flow of graphs whose structural evolution raises fundamental questions. When no size constraints are imposed, under what conditions can we guarantee non-trivial long-term behavior? How do we rigorously characterize structural transitions between consecutive states? We address the first question by introducing the notion of sustainability, a framework for determining whether an iterative process admits periodic graph sequences. We propose and analyze a graph generator that illustrates the theoretical challenges inherent to this concept. For the second question, we develop a metric quantifying structural change across successive iterations. Together, these contributions open new avenues of inquiry and reinforce the interplay between complex network analysis and temporal graph theory.
FrançaisLaBRI 178 -
12:4513:45
La performance des systèmes de mobilité ne repose plus uniquement sur la technologie. IA, automatisation, hypervision, nouveaux usages : partout, le facteur humain s’impose comme un déterminant clé de la sécurité, de l’acceptabilité et de l’efficacité des services de transport.
Lors de cette animation scientifique, le CATIE (Centre Aquitain des Technologies de l’information et Electronique) https://www.catie.fr/language/fr/accueil/ proposera un retour d’expérience concret sur la manière d’intégrer le facteur humain dès la conception de systèmes et services de mobilité. À partir de cas réels menés dans les secteurs aéronautique, ferroviaire et routier, nous explorerons comment l’analyse des comportements, des états physiologiques et des interactions humain–système permet de dépasser les limites des approches purement technologiques.
Cette session s’adresse aux chercheurs, doctorants et acteurs de l’innovation souhaitant mieux comprendre comment le facteur humain peut être intégré de manière opérationnelle dans la conception des mobilités de demain.
Intervenant : Florian Larrue, Responsable Département Systèmes Centrés sur l'Humain – CATIE
Visio -
12:3013:00
Anna Loeff, chargée de projet Données de recherche, Service transversal de Soutien à la recherche (DIRDOC), anime de webinaire.
Présentation de l’entrepôt national pluridisciplinaire Recherche Data Gouv, de l’espace institutionnel de l’université de Bordeaux, ainsi que le détail de la procédure pour déposer et valoriser vos données.Ce format court (30 min) a pour objectif de sensibiliser aux enjeux et pratiques de la science ouverte toute la communauté universitaire de Bordeaux, et particulièrement les chercheurs et doctorants.
-
14:0015:00
In automated learning problems, the task is to find a model that maps given inputs to their corresponding outputs as accurately as possible. Over the past 30 years, machine learning and deep learning have achieved tremendous success in solving this type of problem. However, while the resulting models can be used to make predictions, they offer limited interpretability, i.e. they provide little insight into *how* they solve the problem. For instance, if we train a recurrent neural network to predict whether a sequence of events will lead to a crash, the model cannot describe the sequences of events that lead to crashes and provide solutions to remediate the issue. Program synthesis is a framework for solving learning problems with models that are programs in a domain-specific language, which allows creating interpretable models in the domain of the problem.
In this talk, I will present an enumerative approach for learning Linear Temporal Logic (LTL) formulas from data. I will begin with an introduction to program synthesis, using examples from program de-obfuscation and anomaly explanation. In the second part, I will present the main techniques used in our algorithm: observational equivalence and domination, two pruning techniques used to reduce the search space, and a connection between LTL Learning and the Boolean Set Cover problem. Finally, I will discuss the engineering choices we made when implementing this algorithm in Bolt, an open-source tool available at https://github.com/SynthesisLab/Bolt.
This talk is based on joint work with Nathanaël Fijalkow, Théo Matricon, Baptiste Mouillon and Pierre Vandenhove.
FrançaisLaBRI