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.

Galaxy And Mass Assembly (GAMA): stellar mass growth of spiral galaxies in the cosmic web

Abstract: We look for correlated changes in stellar mass and star formation rate (SFR) along filaments in the cosmic web by examining the stellar masses and UV-derived SFRs of 1799 ungrouped and unpaired spiral galaxies that reside in filaments. We devise multiple distance metrics to characterize the complex geometry of filaments, and find that galaxies closer to the cylindrical centre of a filament have higher stellar masses than their counterparts near the periphery of filaments, on the edges of voids. In addition, these peripheral spiral galaxies have higher SFRs at a given mass. Complementing our sample of filament spiral galaxies with spiral galaxies in tendrils and voids, we find that the average SFR of these objects in different large-scale environments are similar to each other with the primary discriminant in SFR being stellar mass, in line with previous works. However, the distributions of SFRs are found to vary with large-scale environment. Our results thus suggest a model in which in addition to stellar mass as the primary discriminant, the large-scale environment is imprinted in the SFR as a second-order effect. Furthermore, our detailed results for filament galaxies suggest a model in which gas accretion from voids on to filaments is primarily in an orthogonal direction. Overall, we find our results to be in line with theoretical expectations of the thermodynamic properties of the intergalactic medium in different large-scale environments.

Existence and non-existence results for quasilinear elliptic exterior problems with nonlinear boundary conditions

Abstract: Existence and non-existence results are established for quasilinear elliptic problems with nonlinear boundary conditions and lack of compactness. The proofs combine variational methods with the geometrical feature, due to the competition between the different growths of the non-linearities.

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 to video segmentation that is capable of leveraging information present in unlabeled data in order to improve semantic estimates. Our model combines a convolutional architecture and a spatio-temporal transformer recurrent layer that are 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 extensive experiments in the challenging CityScapes and Camvid datasets, and based on 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.