Hal - Université Grenoble Alpes
Not a member yet
268192 research outputs found
Sort by
Whistled Languages and Whistled Speech
No full textInternational audienceWhistled languages are secondary speech codes based on whistles anchored to, and derived from, a spoken language thanks to a relation of acoustic iconicity. Whistled languages essentially consist of a complementary/auxiliary modality of speech mostly used for distance communication that is mainly used in language communities maintaining a traditional lifestyle. Indeed, whistles propagate well in natural surroundings. This special speech practice is often called ‘whistled speech’ because its productions maintain and augment some acoustical features of spoken speech while degrading others, to adapt spoken sentences into a simple melodic line modulated in amplitude and frequency
ProvSQL: A General System for Keeping Track of the Provenance and Probability of Data
No full textInternational audienceWe present the data model, design choices, and performance of ProvSQL, a general and easy-to-deploy provenance tracking and probabilistic database system implemented as a PostgreSQL extension. ProvSQL’s data and query models closely reflect that of a large core of SQL, including multiset semantics, the full relational algebra, and aggregation. A key part of its implementation relies on generic provenance circuits stored in memory-mapped files. We propose benchmarks to measure the overhead of provenance and probabilistic evaluation and demonstrate its scalability and competitiveness with respect to other state-of-the-art systems
Definition and characterization of refueling equilibrium for batch-operated molten salt reactors
No full textInternational audienceThis paper presents a definition of the refueling equilibrium for batch-operated molten salt reactors (MSR). Batch-operated MSR are molten salt reactors for which the refueling option considered is not continuous but batch-wise. In the present work, the molten salt reactor considered is the ARAMIS-A burner reactor, developed within the ISAC (Innovative System for Actinide Conversion) project. This project involves the main contributors of the French nuclear field: CEA, CNRS, EDF, Framatome and Orano. The study of the equilibrium cycle is relevant to evaluate a certain loading strategy for a specific reactor design, notably in terms of fuel fabrication, operating costs, fuel cycle analysis and waste management. Refueling equilibrium is reached when, for a fixed feed-salt, full end of cycle (EOC) in-salt atomic compositions do not vary one from another. In this paper, it is demonstrated that equilibrium compositions depend only on feed-salt properties and a direct method for equilibrium computation is studied. Last, data-sets of equilibrium composition are generated and deep learning models (multilayer perceptrons, a.k.a MLP) are trained so as to provide satisfactory estimation of the equilibrium associated to the fuel composition in the feed-salt. Such MLP models provide precise estimations for dynamic fuel cycle simulation analysis
Fast in-place accumulation
No full textInternational audienceThis paper deals with simultaneously fast and in-place algorithms for formulae where the result has to be linearly accumulated: some output variables are also input variables, linked by a linear dependency. Fundamental examples include the in-place accumulated multiplication of polynomials or matrices, C += AB (that is with only O(1) extra space). The difficulty is to combine in-place computations with fast algorithms: those usually come at the expense of (potentially large) extra temporary space, but with accumulation the output variables are not even available to store intermediate values. We first propose a novel automatic design of fast and in-place accumulating algorithms for any bilinear formulae (and thus for polynomial and matrix multiplication) and then extend it to any linear accumulation of a collection of functions. For this, we relax the in-place model to any algorithm allowed to modify its inputs, provided that those are restored to their initial state afterwards. This allows us to ultimately derive unprecedented in-place accumulating algorithms for fast polynomial multiplications and for Strassen-like matrix multiplications.We then consider the simultaneously fast and in-place computation of the Euclidean polynomial modular remainder R(X) ≡ A(X) mod B(X). Fast algorithms for this usually also come at the expense of a linear amount of extra temporary space. In particular, they require one to first compute and store the whole quotient Q(X) such that A = BQ+R. We here propose an *in-place* algorithm to compute the remainder only. If A and B have respective degree m+n and n, and M(k) denotes the complexity of a (not-in-place) algorithm to multiply two degree-k polynomials, our algorithm uses at most O((n/m) M(m) log(m)) arithmetic operations. In this particular case this is a factor log(n) more than the not-in-place algorithm. But if M(n) = Θ(n^{1+ε}) for some ε>0, then our algorithms do match the not-in-place complexity bound of O((n/m) M(m)). We also propose variants that compute – still in-place and with the same kind of complexity bounds – the over-place remainder A(X) ≡ A(X) mod B(X), the accumulated remainder R(X) += A(X) mod B(X) and the accumulated modular multiplication R(X) += A(X)C(X) mod B(X), that is multiplication in a polynomial extension of a finite field.To achieve this, we develop techniques for Toeplitz matrix operations, for generalized convolutions, short product and power series division and remainder whose output is also part of the input
Using thin AZO layers coupled with SiNx:H as a way to decrease Indium consumption in SHJ cells and modules
No full textInternational audienc
Optimal sub-Gaussian variance proxy for truncated Gaussian and exponential random variables
No full textInternational audienceThis paper establishes the optimal sub-Gaussian variance proxy for truncated Gaussian and truncated exponential random variables. The proofs are based initially on reducing each distribution to their standardized versions. Geometrically, for the normal distribution, our argument consists of fitting a parabola to another parabola-looking function, which emerges from its moment generating function. For the exponential case, we show that the optimal variance proxy is the unique solution to a pair of equations and then provide this solution explicitly. Moreover, we demonstrate that truncated Gaussian variables exhibit strict sub-Gaussian behavior if and only if they are symmetric, meaning their truncation is symmetric with respect to the mean. Conversely, truncated exponential variables are shown to never exhibit strict sub-Gaussianity
