Zobrazeno 1 - 10
of 1 029
pro vyhledávání: '"P. Grangier"'
Momentum based optimizers are central to a wide range of machine learning applications. These typically rely on an Exponential Moving Average (EMA) of gradients, which decays exponentially the present contribution of older gradients. This accounts fo
Externí odkaz:
http://arxiv.org/abs/2409.03137
Autor:
Grangier, Philippe, Auffeves, Alexia, Farouki, Nayla, Bossche, Mathias Van Den, Ezratty, Olivier
The purpose of this article is to provide a novel approach and justification of the idea that classical physics and quantum physics can neither function nor even be conceived one without the other - in line with ideas attributed to e.g. Niels Bohr or
Externí odkaz:
http://arxiv.org/abs/2406.05169
Autor:
Grangier, Philippe
In the foundations of quantum mechanics, the Kolmogorovian Censorship (KC) stipulates that quantum probabilities can be identified with classical, Kolmogorovian probabilities when considering a specified measurement context. Then in any given measure
Externí odkaz:
http://arxiv.org/abs/2405.03184
Autor:
Piétri, Yoann, Schiavon, Matteo, Acosta, Valentina Marulanda, Gouraud, Baptiste, Vidarte, Luis Trigo, Grangier, Philippe, Rhouni, Amine, Diamanti, Eleni
Quantum Key Distribution (QKD) enables secret key exchange between two remote parties with information-theoretic security rooted in the laws of quantum physics. Encoding key information in continuous variables (CV), such as the values of quadrature c
Externí odkaz:
http://arxiv.org/abs/2404.18637
Large language models have emerged as a versatile tool but are challenging to apply to tasks lacking large inference budgets and large in-domain training sets. This work formalizes these constraints and distinguishes four important variables: the pre
Externí odkaz:
http://arxiv.org/abs/2402.01093
Large language models are trained on massive scrapes of the web, which are often unstructured, noisy, and poorly phrased. Current scaling laws show that learning from such data requires an abundance of both compute and data, which grows with the size
Externí odkaz:
http://arxiv.org/abs/2401.16380
Large neural networks pretrained on web-scale corpora are central to modern machine learning. In this paradigm, the distribution of the large, heterogeneous pretraining data rarely matches that of the application domain. This work considers modifying
Externí odkaz:
http://arxiv.org/abs/2311.11973
Large, pre-trained models are problematic to use in resource constrained applications. Fortunately, task-aware structured pruning methods offer a solution. These approaches reduce model size by dropping structural units like layers and attention head
Externí odkaz:
http://arxiv.org/abs/2311.06382
Autor:
Piétri, Yoann, Vidarte, Luis Trigo, Schiavon, Matteo, Vivien, Laurent, Grangier, Philippe, Rhouni, Amine, Diamanti, Eleni
Quantum Key Distribution (QKD) is a prominent application in the field of quantum cryptography providing information-theoretic security for secret key exchange. The implementation of QKD systems on photonic integrated circuits (PICs) can reduce the s
Externí odkaz:
http://arxiv.org/abs/2311.03978
We argue that a clear view on quantum mechanics is obtained by considering that the unicity of the macroscopic world is a fundamental postulate of physics, rather than an issue that must be mathematically justified or demonstrated. This postulate all
Externí odkaz:
http://arxiv.org/abs/2310.06099