Portail HAL Ensta
Not a member yet
11080 research outputs found
Sort by
Acoustic source localization in underwater environment using set methods
No full textInternational audienceThe study of underwater acoustic wave propagation provides solutions to localization and underwater navigation problems. In these cases, simulation can be a powerful tool for a better understanding of acoustic propagation. These simulations are based on models that rely on simplifying assumptions allowing the numerical resolution. Simulation is also used to solve more specific problems in underwater environments. For instance, acoustic source localization using receivers in an underwater scene is still a challenging problem and has both civil and military applications. Classical methods are based on the use of acoustic receiver arrays placed in the environment. Assuming a normal modes model for the propagation, collected data are then processed, for example, by singular value decomposition or matched field processing based approach, which provides probabilistic results. The proposed approach to solve this problem is to use set methods. This method allows enclosing all source positions compatible with the recorded hydrophone signal. In addition, possible sets for source position compatible with each receiver can be intersected to increase the certainty of the source location. Besides requiring a good knowledge of the scene, this method requires simulating the acoustic propagation as well as possible to correctly solve this localization problem
Computing weakly singular and near-singular integrals over curved boundary elements
No full textInternational audienc
La place des ODD dans les programmes de formation des ingénieurs : perceptions croisées des enseignants et des élèves ingénieurs
No full textInternational audienc
GRIMGEP: Learning Progress for Robust Goal Sampling in Visual Deep Reinforcement Learning
No full textInternational audienceDesigning agents, capable of learning autonomously a wide range of skills is critical in order to increase the scope of reinforcement learning. It will both increase the diversity of learned skills and reduce the burden of manually designing reward functions for each skill. Self-supervised agents, setting their own goals, and trying to maximize the diversity of those goals have shown great promise towards this end. However, a currently known limitation of agents trying to maximize the diversity of sampled goals is that they tend to get attracted to noise or more generally to parts of the environments that cannot be controlled (distractors). When agents have access to predefined goal features or expert knowledge, absolute Learning Progress (ALP) provides a way to distinguish between regions that can be controlled and those that cannot. However, those methods often fall short when the agents are only provided with raw sensory inputs such as images. In this work we extend those concepts to unsupervised image-based goal exploration. We propose a framework that allows agents to autonomously identify and ignore noisy distracting regions while searching for novelty in the learnable regions to both improve overall performance and avoid catastrophic forgetting. Our framework can be combined with any stateof- the-art novelty seeking goal exploration approaches. We construct a rich 3D image based environment with distractors. Experiments on this environment show that agents using our framework successfully identify interesting regions of the environment, resulting in drastically improved performances. The source code is available at https://sites.google.com/view/grimgep
Language and Culture Internalisation for Human-Like Autotelic AI
No full textInternational audienceBuilding autonomous agents able to grow open-ended repertoires of skills across their lives is a fundamental goal of artificial intelligence (AI). A promising developmental approach recommends the design of intrinsically motivated agents that learn new skills by generating and pursuing their own goals - autotelic agents. But despite recent progress, existing algorithms still show serious limitations in terms of goal diversity, exploration, generalisation or skill composition. This perspective calls for the immersion of autotelic agents into rich socio-cultural worlds, an immensely important attribute of our environment that shapes human cognition but is mostly omitted in modern AI. Inspired by the seminal work of Vygotsky, we propose Vygotskian autotelic agents - agents able to internalise their interactions with others and turn them into cognitive tools. We focus on language and show how its structure and informational content may support the development of new cognitive functions in artificial agents as it does in humans. We justify the approach by uncovering several examples of new artificial cognitive functions emerging from interactions between language and embodiment in recent works at the intersection of deep reinforcement learning and natural language processing. Looking forward, we highlight future opportunities and challenges for Vygotskian Autotelic AI research, including the use of language models as cultural models supporting artificial cognitive development
Inverse Scattering Theory and Transmission Eigenvalues: Second Edition
No full textInternational audienceIn the first edition of this book, we discussed methods for determining the support of an inhomogeneous media from measured far field data as well as an extensive study of the central role played by the transmission eigenvalue problem in the mathematical development of these methods. In particular, we introduced the generalized linear sampling method (GLSM) and showed that this method provides a mathematical explanation of why it is permissible to use Tikhonov regularization to obtain an approximate solution of the far field equation associated with the linear sampling method.In the five years since the first edition of our book appeared, there has been considerable progress in both the development of GLSM as well as the theory of transmission eigenvalues. In this second edition, in addition to correcting typos in the first edition, we have added several highlights taken from these new developments. In particular, we have included new chapters on 1) the use of modified background media in the nondestructive testing of materials and in particular methods for determining the modified transmission eigenvalues that arise in such applications from measured far field data, 2) a study of a subset of transmission eigenvalues, called non-scattering wave numbers, through the use of techniques taken from the theory of free boundary value problems for elliptic partial differential equations and 3) the duality between scattering poles and transmission eigenvalues which, in addition to their intrinsic mathematical interest, leads to new methods for the numerical computation of scattering poles
SLRL: A Simple Least Remaining Lifetime File Eviction policy for HPC multi-tier storage systems
No full textInternational audienceHPC systems are composed of multiple tiers of storage, from the top high performance tier (high speed SSDs) to the bottom capacitive one (tapes). File placement in such architecture is managed through prefetchers (bottom-up) and eviction policies (top-down). Most state-of-the-art work focus on the former while using algorithm flavors of LRU, LFU and FIFO for the latter. LRU was for long considered the best choice. However, recent studies has shown that the simplicity of FIFO could make it more scalable than LRU because of metadata management, and thus more adequate in several cases. In this paper, we propose a new eviction policy based on predicted files lifetimes. It is comparable to FIFO in terms of metadata overhead and simplicity (thus scalability), while giving a hit ratio comparable to LRU (or even 10% better for some tested traces). We also propose a naive multi-tier heterogeneous storage simulator implementation to evaluate such policies
Estimation of the Influence of a Noisy Environment on the Binary Decision Strategy in a Quantum Illumination Radar
No full textInternational audienceA quantum illumination radar uses quantum entanglement to enhance photodetection sensitivity. The entanglement is quickly destroyed by the decoherence in an environment, although the sensitivity enhancement could survive thanks to quantum correlations beyond the entanglement. These quantum correlations are quantified by the quantum discord. Here, we use a toy model with an amplitude damping channel and Lloyd's binary decision strategy to highlight the possible role of these correlations from the perspective of a quantum radar
