Romanian Academy

About: Romanian Academy is a archive organization based out in Bucharest, Romania. It is known for research contribution in the topics: Population & Nonlinear system. The organization has 3662 authors who have published 10491 publications receiving 146447 citations. The organization is also known as: Academia Română & Societatea Literară Română.

HYBRID γ DORADUS-δ SCUTI PULSATORS: NEW INSIGHTS INTO THE PHYSICS OF THE OSCILLATIONS FROM KEPLER OBSERVATIONS

Abstract: Observations of the pulsations of stars can be used to infer their interior structure and test theoretical models. The main-sequence γ Doradus (Dor) and δ Scuti (Sct) stars with masses 1.2-2.5 M ☉ are particularly useful for these studies. The γ Dor stars pulsate in high-order g-modes with periods of order 1 day, driven by convective blocking at the base of their envelope convection zone. The δ Sct stars pulsate in low-order g- and p-modes with periods of order 2 hr, driven by the κ mechanism operating in the He II ionization zone. Theory predicts an overlap region in the Hertzsprung-Russell diagram between instability regions, where hybrid stars pulsating in both types of modes should exist. The two types of modes with properties governed by different portions of the stellar interior provide complementary model constraints. Among the known γ Dor and δ Sct stars, only four have been confirmed as hybrids. Now, analysis of combined Quarter 0 and Quarter 1 Kepler data for hundreds of variable stars shows that the frequency spectra are so rich that there are practically no pure δ Sct or γ Dor pulsators, i.e., essentially all of the stars show frequencies in both the δ Sct and the γ Dor frequency range. A new observational classification scheme is proposed that takes into account the amplitude as well as the frequency and is applied to categorize 234 stars as δ Sct, γ Dor, δ Sct/γ Dor or γ Dor/δ Sct hybrids.

Semantic Video Segmentation by Gated Recurrent Flow Propagation

Abstract: Semantic video segmentation is challenging due to the sheer amount of data that needs to be processed and labeled in order to construct accurate models. In this paper we present a deep, end-to-end trainable methodology for video segmentation that is capable of leveraging the information present in unlabeled data, besides sparsely labeled frames, in order to improve semantic estimates. Our model combines a convolutional architecture and a spatiotemporal transformer recurrent layer that is able to temporally propagate labeling information by means of optical flow, adaptively gated based on its locally estimated uncertainty. The flow, the recognition and the gated temporal propagation modules can be trained jointly, end-to-end. The temporal, gated recurrent flow propagation component of our model can be plugged into any static semantic segmentation architecture and turn it into a weakly supervised video processing one. Our experiments in the challenging CityScapes and Camvid datasets, and for multiple deep architectures, indicate that the resulting model can leverage unlabeled temporal frames, next to a labeled one, in order to improve both the video segmentation accuracy and the consistency of its temporal labeling, at no additional annotation cost and with little extra computation.

Ground state solutions of scalar field fractional Schrödinger equations

Abstract: In this paper, we study the existence of multiple ground state solutions for a class of parametric fractional Schrodinger equations whose simplest prototype is (−�) s u + V (x)u = λ f (x, u) in R n , where n > 2 ,( −�) s stands for the fractional Laplace operator of order s ∈ (0,1) ,a ndλ is a positive real parameter. The nonlinear term f is assumed to have a superlinear behaviour at the origin and a sublinear decay at infinity. By using variational methods, we establish the existence of a suitable range of positive eigenvalues for which the problem admits at least two nontrivial solutions in a suitable weighted fractional Sobolev space. Mathematics Subject Classification 35A15 · 35S15 · 49J35 · 45G05 · 47G20

Small universal spiking neural P systems.

Abstract: In search for small universal computing devices of various types, we consider here the case of spiking neural P systems (SN P systems), in two variants: as devices that compute functions and as devices that generate sets of numbers. We start with the first case and we produce a universal spiking neural P system with 84 neurons. If a slight generalization of the used rules is adopted, namely, we allow rules for producing simultaneously several spikes, then a considerable reduction, to 49 neurons, is obtained. For SN P systems used as generators of sets of numbers, we find a universal system with restricted rules having 76 neurons and one with extended rules having 50 neurons.