Extractive summarization using complex networks and syntactic dependency

The realization that statistical physics methods can be applied to analyze written texts represented as complex networks has led to several developments in natural language processing, including automatic summarization and evaluation of machine translation. Most importantly, so far only a few metrics of complex networks have been used and therefore there is ample opportunity to enhance the statistics-based methods as new measures of network topology and dynamics are created. In this paper, we employ for the first time the metrics betweenness, vulnerability and diversity to analyze written texts in Brazilian Portuguese. Using strategies based on diversity metrics, a better performance in automatic summarization is achieved in comparison to previous work employing complex networks. With an optimized method the Rouge score (an automatic evaluation method used in summarization) was 0.5089, which is the best value ever achieved for an extractive summarizer with statistical methods based on complex networks for Brazilian Portuguese. Furthermore, the diversity metric can detect keywords with high precision, which is why we believe it is suitable to produce good summaries. It is also shown that incorporating linguistic knowledge through a syntactic parser does enhance the performance of the automatic summarizers, as expected, but the increase in the Rouge score is only minor. These results reinforce the suitability of complex network methods for improving automatic summarizers in particular, and treating text in general. (C) 2011 Elsevier B.V. All rights reserved.

Diet of Astyanax paranae (Characidae) in streams with different riparian land covers in the Passa-Cinco River basin, southeastern Brazil

An analysis of the diet of Astyanax paranae Eigenmann, 1914 in nine streams located in the Passa-Cinco River basin (upper Parana River system) was performed to investigate the feeding habits of this species, check for possible spatial variations in diet and to investigate the influence of riparian vegetation in the composition of the diet. Stomach contents of 243 specimens were analyzed by the methods of relative frequency of occurrence and volume, and the diet was characterized by the alimentary index (AI(i)). The species showed insectivorous feeding habits, with a predominance of terrestrial and aquatic insects in the diet, varying by location. In most streams, resources of allochthonous origin were the most consumed. The participation of aquatic insects and terrestrial plants were high in most streams, while terrestrial insects and invertebrates were highest in streams with a greater presence of riparian forest. The two streams located draining pasture fields were the only places were A. paranae consumed algae and macrophyte fragments. These results were corroborated by the analysis of similarity (ANOSIM): the descriptor "percentage of riparian forest" was the highest environmental influence on the diet of A. paranae. The study shows that riparian forest percentage on the stream reach determines the species diet composition, but A. paranae is also able to gather enough food resources in a variety of severely degraded environments.

Physical-chemical and microbiological changes in Cerrado Soil under differing sugarcane harvest management systems

Background: Sugarcane cultivation plays an important role in Brazilian economy, and it is expanding fast, mainly due to the increasing demand for ethanol production. In order to understand the impact of sugarcane cultivation and management, we studied sugarcane under different management regimes (pre-harvest burn and mechanical, unburnt harvest, or green cane), next to a control treatment with native vegetation. The soil bacterial community structure (including an evaluation of the diversity of the ammonia oxidizing (amoA) and denitrifying (nirK) genes), greenhouse gas flow and several soil physicochemical properties were evaluated. Results: Our results indicate that sugarcane cultivation in this region resulted in changes in several soil properties. Moreover, such changes are reflected in the soil microbiota. No significant influence of soil management on greenhouse gas fluxes was found. However, we did find a relationship between the biological changes and the dynamics of soil nutrients. In particular, the burnt cane and green cane treatments had distinct modifications. There were significant differences in the structure of the total bacterial, the ammonia oxidizing and the denitrifying bacterial communities, being that these groups responded differently to the changes in the soil. A combination of physical and chemical factors was correlated to the changes in the structures of the total bacterial communities of the soil. The changes in the structures of the functional groups follow a different pattern than the physicochemical variables. The latter might indicate a strong influence of interactions among different bacterial groups in the N cycle, emphasizing the importance of biological factors in the structuring of these communities. Conclusion: Sugarcane land use significantly impacted the structure of total selected soil bacterial communities and ammonia oxidizing and denitrifier gene diversities in a Cerrado field site in Central Brazil. A high impact of land use was observed in soil under the common burnt cane management. The green cane soil also presented different profiles compared to the control soil, but to at a lesser degree.

Semi-supervised learning guided by the modularity measure in complex networks

Semi-supervised learning techniques have gained increasing attention in the machine learning community, as a result of two main factors: (1) the available data is exponentially increasing; (2) the task of data labeling is cumbersome and expensive, involving human experts in the process. In this paper, we propose a network-based semi-supervised learning method inspired by the modularity greedy algorithm, which was originally applied for unsupervised learning. Changes have been made in the process of modularity maximization in a way to adapt the model to propagate labels throughout the network. Furthermore, a network reduction technique is introduced, as well as an extensive analysis of its impact on the network. Computer simulations are performed for artificial and real-world databases, providing a numerical quantitative basis for the performance of the proposed method.

Effective number of accessed nodes in complex networks

The measurement called accessibility has been proposed as a means to quantify the efficiency of the communication between nodes in complex networks. This article reports results regarding the properties of accessibility, including its relationship with the average minimal time to visit all nodes reachable after h steps along a random walk starting from a source, as well as the number of nodes that are visited after a finite period of time. We characterize the relationship between accessibility and the average number of walks required in order to visit all reachable nodes (the exploration time), conjecture that the maximum accessibility implies the minimal exploration time, and confirm the relationship between the accessibility values and the number of nodes visited after a basic time unit. The latter relationship is investigated with respect to three types of dynamics: traditional random walks, self-avoiding random walks, and preferential random walks.

NATURAL REGENERATION IN ABANDONED FIELDS FOLLOWING INTENSIVE AGRICULTURAL LAND USE IN AN ATLANTIC FOREST ISLAND, BRAZIL

The time required to regrowth a forest in degraded areas depends on how the forest is removed and on the type of land use following removal. Natural regeneration was studied in abandoned old fields after intensive agricultural land use in areas originally covered by Brazilian Atlantic Forests of the Anchieta Island, Brazil in order to understand how plant communities reassemble following human disturbances as well as to determine suitable strategies of forest restoration. The fields were classified into three vegetation types according to the dominant plant species in: 1) Miconia albicans (Sw.) Triana (Melastomataceae) fields, 2) Dicranopteris flexuosa (Schrader) Underw. (Gleicheniaceae) thickets, and 3) Gleichenella pectinata (Willd.) Ching. (Gleicheniaceae) thickets. Both composition and structure of natural regeneration were compared among the three dominant vegetation types by establishing randomly three plots of 1 x 3 m in five sites of the island. A gradient in composition and abundance of species in natural regeneration could be observed along vegetation types from Dicranopteris fern thickets to Miconia fields. The gradient did not accurately follow the pattern of spatial distribution of the three dominant vegetation types in the island regarding their proximity of the remnant forests. A complex association of biotic and abiotic factors seems to be affecting the seedling recruitment and establishment in the study plots. The lowest plant regeneration found in Dicranopteris and Gleichenella thickets suggests that the ferns inhibit the recruitment of woody and herbaceous species. Otherwise, we could not distinguish different patterns of tree regeneration among the three vegetation types. Our results showed that forest recovery following severe anthropogenic disturbances is not direct, predictable or even achievable on its own. Appropriated actions and methods such as fern removal, planting ground covers, and enrichment planting with tree species were suggested in order to restore the natural forest regeneration process in the abandoned old fields.

Stochastic competitive learning in complex networks

Competitive learning is an important machine learning approach which is widely employed in artificial neural networks. In this paper, we present a rigorous definition of a new type of competitive learning scheme realized on large-scale networks. The model consists of several particles walking within the network and competing with each other to occupy as many nodes as possible, while attempting to reject intruder particles. The particle's walking rule is composed of a stochastic combination of random and preferential movements. The model has been applied to solve community detection and data clustering problems. Computer simulations reveal that the proposed technique presents high precision of community and cluster detections, as well as low computational complexity. Moreover, we have developed an efficient method for estimating the most likely number of clusters by using an evaluator index that monitors the information generated by the competition process itself. We hope this paper will provide an alternative way to the study of competitive learning.

Phylodynamics and movement of Phycodnaviruses among aquatic environments

Phycodnaviruses have a significant role in modulating the dynamics of phytoplankton, thereby influencing community structure and succession, nutrient cycles and potentially atmospheric composition because phytoplankton fix about half the carbon dioxide (CO2) on the planet, and some algae release dimethylsulphoniopropionate when lysed by viruses. Despite their ecological importance and widespread distribution, relatively little is known about the evolutionary history, phylogenetic relationships and phylodynamics of the Phycodnaviruses from freshwater environments. Herein we provide novel data on Phycodnaviruses from the largest river system on earth-the Amazon Basin-that were compared with samples from different aquatic systems from several places around the world. Based on phylogenetic inference using DNA polymerase (pol) sequences we show the presence of distinct populations of Phycodnaviridae. Preliminary coarse-grained phylodynamics and phylogeographic inferences revealed a complex dynamics characterized by long-term fluctuations in viral population sizes, with a remarkable worldwide reduction of the effective population around 400 thousand years before the present (KYBP), followed by a recovery near to the present time. Moreover, we present evidence for significant viral gene flow between freshwater environments, but crucially almost none between freshwater and marine environments. The ISME Journal (2012) 6, 237-247; doi: 10.1038/ismej.2011.93; published online 28 July 2011

The genus Nanopterodectes Mironov, 2009 (Acari, Proctophyllodidae), with descriptions of three new species from antbirds (Passeriformes: Thamnophilidae) in Brazil

Three new species of the recently discovered, and hitherto monotypic, feather mite genus Nanopterodectes Mironov, 2009 are described: N. acutirostris n. sp. from Stymphalornis acutirostris Bornschein, Reinert & Teixeira, N. mentalis n. sp. from Dysithamnus mentalis (Temminck), and N. leucopterus n. sp. from Pyriglena leucoptera (Vieillot). This feather mite genus is currently restricted to passerine birds of the Neotropical family Thamnophilidae in Brazil. A key to the known species of Nanopterodectes is presented for both sexes.

Shifts in phylogenetic diversity of archaeal communities in mangrove sediments at different sites and depths in southeastern Brazil

This study focused on the structure and composition of archaeal communities in sediments of tropical mangroves in order to obtain sufficient insight into two Brazilian sites from different locations (one pristine and another located in an urban area) and at different depth levels from the surface. Terminal restriction fragment length polymorphism (T-RFLP) of PCR-amplified 16S rRNA gene fragments was used to scan the archaeal community structure, and 16S rRNA gene clone libraries were used to determine the community composition. Redundancy analysis of T-RFLP patterns revealed differences in archaeal community structure according to location, depth and soil attributes. Parameters such as pH, organic matter, potassium and magnesium presented significant correlation with general community structure. Furthermore, phylogenetic analysis revealed a community composition distributed differently according to depth where, in shallow samples, 74.3% of sequences were affiliated with Euryarchaeota and 25.7% were shared between Crenarchaeota and Thaumarchaeota, while for the deeper samples, 24.3% of the sequences were affiliated with Euryarchaeota and 75.7% with Crenarchaeota and Thaumarchaeota. Archaeal diversity measurements based on 16S rRNA gene clone libraries decreased with increasing depth and there was a greater difference between depths (<18% of sequences shared) than sites (>25% of sequences shared). Taken together, our findings indicate that mangrove ecosystems support a diverse archaeal community; it might possibly be involved in nutrient cycles and are affected by sediment properties, depth and distinct locations. (C) 2012 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved.

Evaluation of impacts of climate change and local stressors on the biotechnological potential of marine macroalgae: a brief theoretical discussion of likely scenarios

Climate change can be associated with variations in the frequency and intensity of extreme temperatures and precipitation events on the local and regional scales. Along coastal areas, flooding associated with increased occupation has seriously impacted products and services generated by marine life, in particular the biotechnological potential that macroalgae hold. Therefore, this paper analyzes the available information on the taxonomy, ecology and physiology of macroalgae and discusses the impacts of climate change and local stress on the biotechnological potential of Brazilian macroalgae. Based on data compiled from a series of floristic and ecological works, we note the disappearance in some Brazilian regions of major groups of biotechnological interest. In some cases, the introduction of exotic species has been documented, as well as expansion of the distribution range of economically important species. We also verify an increase in the similarities between the Brazilian phycogeographic provinces, although they still remain different. It is possible that these changes have resulted from the warming of South Atlantic water, as observed for its surface in southeastern Brazilian, mainly during the winter. However, unplanned urbanization of coastal areas can also produce similar biodiversity losses, which requires efforts to generate long-term temporal data on the composition, community structure and physiology of macroalgae.

Changes in phytoplankton composition in response to tides, wind-induced mixing conditions, and freshwater outflows in an urbanised estuarine complex

Recent reports have shown an increase in potentially harmful phytoplankton in Santos bay (Southeastern Brazilian Coast), located in a highly urbanised estuarine complex. Prediction of blooms is, thus, essential but the phytoplankton community structure in very dynamic regions is difficult to determine. In the present work, we discriminate bloom forming microphytoplankton dominance and their relationship to physical and meteorological variables to look for patterns observed in different tides and seasons. Comparing 8 distinct situations, we found five scenarios of dominance that could be related to winds, tides and rainfall: i) Surfers, diatoms occurring during high surf zone energies; ii) Sinkers, represented by larger celled diatoms during spring tide, after periods of high precipitation rates; iii) Opportunistic mixers, composed of chain forming diatoms with small or elongate cells occurring during neap tides; iv) Local mixers, microplanktonic diatoms and dinoflagellates which occurred throughout the 298 sampling stations; and v) Mixotrophic dinoflagellates, after intense estuarine discharges. Results suggest alterations in the temporal patterns for some bloom-forming species, while others appeared in abundances above safe limits for public health. This approach can also illustrate possible impacts of changes in freshwater discharge in highly urbanised estuaries.

Regional and temporal variability of sinking organic matter in the subtropical northeast Atlantic Ocean: a biomarker diagnosis

Sinking particles through the pelagic ocean have been traditionally considered the most important vehicle by which the biological pump sequesters carbon in the ocean interior. Nevertheless, regional scale variability in particle flux is a major outstanding issue in oceanography. 5 Here, we have studied the regional and temporal variability of total particulate organic matter fluxes, as well as chloropigment and total hydrolyzed amino acid (THAA) compositions and fluxes in the Canary Current region, between 20–30 N, during two contrasting periods: August 2006, characterized by warm and stratified waters, but also intense winds which enhanced eddy development south of the Canary Islands, 10 and February 2007, characterized by colder waters, less stratification and higher productivity. We found that the eddy-field generated south of the Canary Islands enhanced by >2 times particulate organic carbon (POC) export with respect to stations (FF; farfield) outside the eddy-field influence. We also observed flux increases of one order of magnitude in chloropigment and 70% in THAA in the eddy-field relative to FF stations. 15 Principal Components Analysis (PCA) was performed to assess changes in particulate organic matter composition between stations. At eddy-field stations, higher chlorophyll enrichment reflected “fresher” material, while at FF stations a higher proportion of pheophytin indicated greater degradation due to microbes and microzooplankton. PCA also suggests that phytoplankton community structure, particularly the dominance of 20 diatoms versus carbonate-rich plankton, is the major factor influencing the POC export within the eddy field. In February, POC export fluxes were the highest ever reported for this area, reaching values of 15 mmolCm−2 d−1 at 200m depth. Compositional changes in pigments and THAA indicate that the source of sinking particles varies zonally and meridionally and suggest that sinking particles were more degraded at 25 near-coastal stations relative to open ocean stations.

Regional and temporal variability of sinking organic matter in the subtropical northeast Atlantic Ocean: A biomarker diagnosis

[EN] Sinking particles through the pelagic ocean have been traditionally considered the most important vehicle by which the biological pump sequesters carbon in the ocean interior. Nevertheless, regional scale variability in particle flux is a major outstanding issue in oceanography. Here, we have studied the regional and temporal variability of total particulate organic matter fluxes, as well as chloropigment and total hydrolyzed amino acid (THAA) compositions and fluxes in the Canary Current region, between 20?30_ N, during two contrasting periods: August 2006, characterized by warm and stratified waters, but also intense winds which enhanced eddy development south of the Canary Islands, and February 2007, characterized by colder waters, less stratification and higher productivity. We found that the eddyfield generated south of the Canary Islands enhanced by >2 times particulate organic carbon (POC) export with respect to stations (FF; far-field) outside the eddy-field influence. We also observed flux increases of one order of magnitude in chloropigment and 2 times in THAA in the eddy-field relative to FF stations. Principal Components Analysis (PCA) was performed to assess changes in particulate organic matter composition between stations. At eddy-field stations, higher chlorophyll enrichment reflected ?fresher? material, while at FF stations a higher proportion of pheophytin indicated greater degradation due to microbes and microzooplankton. PCA also suggests that phytoplankton community structure, particularly the dominance of diatoms versus carbonate-rich plankton, is the major factor influencing the POC export within the eddy field. In February, POC export POC export within the eddy field. In February, POC export fluxes were the highest ever reported for this area, reaching values of _15 mmolCm?2 d?1 at 200m depth. Compositional changes in pigments and THAA indicate that the source of sinking particles varies zonally and meridionally and suggest that sinking particles were more degraded at near-coastal stations relative to open ocean stations.

Role of slowly settling particles in the ocean carbon cycle

[EN] Here we present results from sediment traps that separate particles as a function of their settling velocity, which were moored in the Canary Current region over a 1.5-year period. This study represents the longest time series using “in situ” particle settling velocity traps to date and are unique in providing year-round estimates. We find that, at least during half of the year in subtropical waters (the largest ocean domain), more than 60% of total particulate organic carbon is contained in slowly settling particles (0.7–11 m d−1). Analyses of organic biomarkers reveal that these particles have the same degradation state, or are even fresher than rapidly sinking particles. Thus, if slowly settling particles dominate the exportable carbon pool, most organic matter would be respired in surface waters, acting as a biological source of CO2 susceptible to exchange with the atmosphere. In the context of climate change, if the predicted changes in phytoplankton community structure occur, slowly settling particles would be favored, affecting the strength of the biological pump in the ocean.