Search for: All records

Fleshy macroalgae dominate the hard bottom, shallow waters along the Western Antarctic Peninsula (WAP). Although there are numerous reports on their ecology, geographic distribution, and to a lesser extent, vertical (depth) distribution in the northern portions of the WAP, much less is known farther south along the central portion of the WAP. Here we provide the first report of the vertical distributions of brown and red fleshy macroalgae in this region based on scuba-derived collections at 14 study sites between southern Anvers Island (64.8°S, 64.4°W) in the north and central Marguerite Bay (68.7°S, 67.5°W) in the south. Although several overstory brown macroalgal species that can be common along the northern WAP includingDesmarestia ancepsandCystosphaera jacquinotiiare mostly absent from the central WAP, the vertical distributions of the brown macroalgaeDesmarestia menziesiiandHimantothallus grandifoliusare similar to the northern WAP even though their percent cover is much lower. Likewise, the vertical distribution of the 14 most widespread red macroalgae, where they occur, mirrored those known from the northern part of the WAP even though macroalgal cover, biomass, and total species richness declined markedly to the south across this region due to increasing sea ice concentrations.

 

Macroalgal forests dominate shallow hard bottom areas along the northern portion of the Western Antarctic Peninsula (WAP). Macroalgal biomass and diversity are known to be dramatically lower in the southern WAP and at similar latitudes around Antarctica, but few reports detail the distributions of macroalgae or associated macroinvertebrates in the central WAP. We used satellite imagery to identify 14 sites differing in sea ice coverage but similar in terms of turbidity along the central WAP. Fleshy macroalgal cover was strongly, negatively correlated with ice concentration, but there was no significant correlation between macroinvertebrate cover and sea ice. Overall community (all organisms) diversity correlated negatively with sea ice concentration and positively with fleshy macroalgal cover, which ranged from around zero at high ice sites to 80% at the lowest ice sites. Nonparametric, multivariate analyses resulted in clustering of macroalgal assemblages across most of the northern sites of the study area, although they differed greatly with respect to macroalgal percent cover and diversity. Analyses of the overall communities resulted in three site clusters corresponding to high, medium, and low fleshy macroalgal cover. At most northern sites, macroalgal cover was similar across depths, but macroalgal and macroinvertebrate distributions suggested increasing effects of ice scour in shallower depths towards the south. Hindcast projections based on correlations of ice and macroalgal cover data suggest that macroalgal cover at many sites could have been varying substantially over the past 40 years. Similarly, based on predicted likely sea ice decreases by 2100, projected increases in macroalgal cover at sites that currently have high ice cover and low macroalgal cover are substantial, often with only a future 15% decrease in sea ice. Such changes would have important ramifications to future benthic communities and to understanding how Antarctic macroalgae may contribute to future blue carbon sequestration. 

This work explores the use of a triplet neural net-work for assessing the similarity of paper textures in a collection of Henri Matisse’s lithographs. The available dataset contains digital photomicrographs of papers in the lithograph collection, consisting of four views: two raking light orientations and both sides of the paper. A triplet neural network is first trained to extract features sensitive to anisotropy, and subsequently used to ensure that all papers in the dataset are in the same orientation and side. Another triplet neural network is then used to extract the texture features that are used to assess paper texture similarity. These results can then be used by art conservators and historians to answer questions of art historical significance, such as artist intent. 

In the context of papers used in the graphic arts, including silver gelatin, inkjet, and wove papers, prior work has studied measures of texture similarity for purposes of classifying such papers. The majority of prior work has been based on classical image processing approaches such as Fourier, wavelet, and fractal analysis. In this work, recent advances in deep learning are used to develop a texture similarity approach for measuring paper texture similarity. Since the available datasets generally lack labels, the convolutional neural network is trained using triplet loss to minimize the feature distance of tiles from the same image while simultaneously maximizing the feature distance of tiles drawn from different images. The approach is tested on three paper texture image databases considered in prior works and the results suggest the proposed approach achieves state-of-the-art performance. 

This paper considers the problem of tracking and predicting dynamical processes with model switching. The classical approach to this problem has been to use an interacting multiple model (IMM) which uses multiple Kalman filters and an auxiliary system to estimate the posterior probability of each model given the observations. More recently, data-driven approaches such as recurrent neural networks (RNNs) have been used for tracking and prediction in a variety of settings. An advantage of data-driven approaches like the RNN is that they can be trained to provide good performance even when the underlying dynamic models are unknown. This paper studies the use of temporal convolutional networks (TCNs) in this setting since TCNs are also data-driven but have certain structural advantages over RNNs. Numerical simulations demonstrate that a TCN matches or exceeds the performance of an IMM and other classical tracking methods in two specific settings with model switching: (i) a Gilbert-Elliott burst noise communication channel that switches between two different modes, each modeled as a linear system, and (ii) a maneuvering target tracking scenario where the target switches between a linear constant velocity mode and a nonlinear coordinated turn mode. In particular, the results show that the TCN tends to identify a mode switch as fast or faster than an IMM and that, in some cases, the TCN can perform almost as well as an omniscient Kalman filter with perfect knowledge of the current mode of the dynamical system. 

This paper studies the “age of information” in a general multi-source multi-hop wireless network with explicit channel contention. Specifically, the scenario considered in this paper assumes that each node in the network is both a source and a monitor of information, that all nodes wish to receive fresh status updates from all other nodes in the network, and that only one node can transmit in each time slot. Lower bounds for peak and average age of information are derived and expressed in terms of fundamental graph properties including the connected domination number. An algorithm to generate near-optimal periodic status update schedules based on sequential optimal flooding is also developed. These schedules are analytically shown to exactly achieve the peak age bound and also achieve the average age bound within an additive gap scaling linearly with the size of the network. Moreover, the results are sufficiently general to apply to any connected network topology. Illustrative numerical examples are presented which serve to verify the analysis for several canonical network topologies of arbitrary size, as well as every connected network with nine or fewer nodes.