École Normale Supérieure

About: École Normale Supérieure is a other organization based out in Paris, Île-de-France, France. It is known for research contribution in the topics: Population & Catalysis. The organization has 68439 authors who have published 99414 publications receiving 3092008 citations.

Time-temperature superposition and dynamic mechanical behavior of atactic polystyrene

Abstract: The dynamic mechanical behavior of monodisperse atactic polystyrene (mol. wt. 98,000) has been measured in the frequency range, 10−5 to 10 Hz and temperature range 359–374K. The time-temperature superposition of the entire data in the frequency range of overlap seems less satisfactory in both the real and imaginary components of the complex shear modulus, G′ and G″, respectively. The lack of adequate superposition becomes pronounced in the tan ϕ (G″/G′) plots. The tan ϕ plots provide a more discerning criteria for the superposition than the G′ or G″ spectra. An analysis based on an earlier model for anelastic deformation shows that of the several changes that may occur in the dynamic mechanical behavior on heating of polystyrene, the predominant ones are both an increase in the size of the microshear domains and the correlations of movement of segments near entanglements. These decrease the contribution to the modulus on heating near Tg so that the time-temperature superposition is vitiated.

The Role of Speech Rhythm in Language Discrimination: Further Tests with a Non-Human Primate.

Abstract: Human newborns discriminate languages from different rhythmic classes, fail to discriminate languages from the same rhythmic class, and fail to discriminate languages when the utterances are played backwards. Recent evidence showing that cotton-top tamarins discriminate Dutch from Japanese, but not when utterances are played backwards, is compatible with the hypothesis that rhythm discrimination is based on a general perceptual mechanism inherited from a primate ancestor. The present study further explores the rhythm hypothesis for language discrimination by testing languages from the same and different rhythmic class. We find that tamarins discriminate Polish from Japanese (different rhythmic classes), fail to discriminate English and Dutch (same rhythmic class), and fail to discriminate backwards utterances from different and same rhythmic classes. These results provide further evidence that language discrimination in tamarins is facilitated by rhythmic differences between languages, and suggest that, in humans, this mechanism is unlikely to have evolved specifically for language.

Leveraging the Present to Anticipate the Future in Videos

Abstract: Anticipating actions before they are executed is crucial for a wide range of practical applications including autonomous driving and the moderation of live video streaming. While most prior work in this area requires partial observation of executed actions, in the paper we focus on anticipating actions seconds before they start. Our proposed approach is the fusion of a purely anticipatory model with a complementary model constrained to reason about the present. In particular, the latter predicts present action and scene attributes, and reasons about how they evolve over time. By doing so, we aim at modeling action anticipation at a more conceptual level than directly predicting future actions. Our model outperforms previously reported methods on the EPIC-KITCHENS and Breakfast datasets.

Sparse coding for machine learning, image processing and computer vision

Abstract: We study in this thesis a particular machine learning approach to represent signals that that consists of modelling data as linear combinations of a few elements from a learned dictionary. It can be viewed as an extension of the classical wavelet framework, whose goal is to design such dictionaries (often orthonormal basis) that are adapted to natural signals. An important success of dictionary learning methods has been their ability to model natural image patches and the performance of image denoising algorithms that it has yielded. We address several open questions related to this framework: How to efficiently optimize the dictionary? How can the model be enriched by adding a structure to the dictionary? Can current image processing tools based on this method be further improved? How should one learn the dictionary when it is used for a different task than signal reconstruction? How can it be used for solving computer vision problems? We answer these questions with a multidisciplinarity approach, using tools from statistical machine learning, convex and stochastic optimization, image and signal processing, computer vision, but also optimization on graphs.