Investigation of multilingual deep neural networks for spoken term detection

The development of high-performance speech processing systems for low-resource languages is a challenging area. One approach to address the lack of resources is to make use of data from multiple languages. A popular direction in recent years is to use bottleneck features, or hybrid systems, trained on multilingual data for speech-to-text (STT) systems. This paper presents an investigation into the application of these multilingual approaches to spoken term detection. Experiments were run using the IARPA Babel limited language pack corpora (∼10 hours/language) with 4 languages for initial multilingual system development and an additional held-out target language. STT gains achieved through using multilingual bottleneck features in a Tandem configuration are shown to also apply to keyword search (KWS). Further improvements in both STT and KWS were observed by incorporating language questions into the Tandem GMM-HMM decision trees for the training set languages. Adapted hybrid systems performed slightly worse on average than the adapted Tandem systems. A language independent acoustic model test on the target language showed that retraining or adapting of the acoustic models to the target language is currently minimally needed to achieve reasonable performance. © 2013 IEEE.

Population differentiation of Procypris rabaudi (Tchang) in the upper reaches of the Yangtze River revealed by mitochondrial DNA and RAPD markers

Procypris rabaudi (Tchang) is a cyprinid fish endermic to middle and upper reaches of the Yangtze River. Besides in main stream and large tributaries, there exists an early matured, small-sized ecological type in a small tributary, Tang River. In this study, mitochondrial DNA cytochrome h (cyt b) gene sequence analysis and randomly amplified polymorphic DNA (RAPD) analysis were performed to investigate the differentiation of the Tang River population from the Mudong reach population of the Yangtze River, with the purpose of conservation and exploitation of this fish. In the 1140 bps of cyt b gene sequence surveyed, 20 sites were found polymorphic, which defined 23 haplotypes. Among them, four haplotypes accounted for 54.4% of all individuals, while population-specific haplotypes occurred in low frequencies. Analysis of molecular variation on cyt b data revealed no significant partition existing between Tang River population and Mudong reach population. Analyses of 132 RAPD loci suggested that genetic variation between populations was significant, though values of different F-ST were not very high. The results revealed low genetic diversity and the beginning of population differentiation, suggesting that Tang River population should be designated as a separate Management Unit.

A gene cluster that regulates both heterocyst differentiation and pattern formation in Anabaena sp strain PCC 7120

Wild-type Anabaena sp. strain PCC 7120, a filamentous nitrogen-fixing cyanobacterium, produces single heterocysts at semi-regular intervals. asr0100 (patU5) and alr0101 (patU3) are homologous to the 5' and 3' portions of patU of Nostoc punctiforme. alr0099 (hetZ) overlaps the 5' end of patU5. hetZ, patU5 and patU3 were all upregulated, or expressed specifically, in proheterocysts and heterocysts. Mutants of hetZ showed delayed or no heterocyst differentiation. In contrast, a patU3 mutation produced a multiple contiguous heterocyst (Mch) phenotype and restored the formation of otherwise lost intercalary heterocysts in a patA background. Decreasing the expression of patU3 greatly increased the frequency of heterocysts in a mini-patS strain. Two promoter regions and two principal, corresponding transcripts were detected in the hetZ-patU5-patU3 region. Transcription of hetZ was upregulated in a hetZ mutant and downregulated in a patU3 mutant. When mutants hetZ::C.K2 and hetZ::Tn5-1087b were nitrogen-deprived, P-hetC-gfp was very weakly expressed, and in hetZ::Tn5-1087b, P-hetR-gfp was relatively strongly expressed in cells that had neither a regular pattern nor altered morphology. We conclude that the hetZ-patU5-patU3 cluster plays an important role in co-ordination of heterocyst differentiation and pattern formation. The presence of homologous clusters in filamentous genera without heterocysts is suggestive of a more general role.

Prediction and elucidation of the population dynamics of Microcystis spp. in Lake Dianchi (China) by means of artificial neural networks

Lake Dianchi is a shallow and turbid lake, located in Southwest China. Since 1985, Lake Dianchi has experienced severe cyanabacterial blooms (dominated by Microcystis spp.). In extreme cases, the algal cell densities have exceeded three billion cells per liter. To predict and elucidate the population dynamics ofMicrocystis spp. in Lake Dianchi, a neural network based model was developed. The correlation coefficient (R 2) between the predicted algal concentrations by the model and the observed values was 0.911. Sensitivity analysis was performed to clarify the algal dynamics to the changes of environmental factors. The results of a sensitivity analysis of the neural network model suggested that small increases in pH could cause significantly reduced algal abundance. Further investigations on raw data showed that the response of Microcystis spp. concentration to pH increase was dependent on algal biomass and pH level. When Microcystis spp. population and pH were moderate or low, the response of Microcystis spp. population would be more likely to be positive in Lake Dianchi; contrarily, Microcystis spp. population in Lake Dianchi would be more likely to show negative response to pH increase when Microcystis spp. population and pH were high. The paper concluded that the extremely high concentration of algal population and high pH could explain the distinctive response of Microcystis spp. population to +1 SD (standard deviation) pH increase in Lake Dianchi. And the paper also elucidated the algal dynamics to changes of other environmental factors. One SD increase of water temperature (WT) had strongest positive relationship with Microcystis spp. biomass. Chemical oxygen demand (COD) and total phosphorus (TP) had strong positive effect on Microcystis spp. abundance while total nitrogen (TN), biological oxygen demand in five days (BOD5), and dissolved oxygen had only weak relationship with Microcystis spp. concentration. And transparency (Tr) had moderate positive relationship with Microcystis spp. concentration.

Modelling algal blooms in Lake Dianchi, China using neural networks

Lake Dianchi is one of the most extensively impacted freshwater lakes by algal blooms. To investigate the response of dominant algal genera, neural networks were applied to model the relationship between water quality parameters and the biomass of four dominant genera (Microcystic spp., Anabaena sp., Quadricauda (Turp.) Breb, Pediastrum Mey) in Dianchi. Results showed that the timing and magnitude of algal blooms of Microcystic spp., nabaena sp., Quadricauda (Turp.) Breb, and Pediastrum Mey in Dianchi could be successfully predicted. The evaluation of environmental factors showed that pH had more significant impact on concentrations of all the four dominant algal genera than the nutrient factors, such as total phosphorus and total nitrogen.

The neural signature of escalating frustration in humans.

Mammalian studies show that frustration is experienced when goal-directed activity is blocked. Despite frustration's strongly negative role in health, aggression and social relationships, the neural mechanisms are not well understood. To address this we developed a task in which participants were blocked from obtaining a reward, an established method of producing frustration. Levels of experienced frustration were parametrically varied by manipulating the participants' motivation to obtain the reward prior to blocking. This was achieved by varying the participants' proximity to a reward and the amount of effort expended in attempting to acquire it. In experiment 1, we confirmed that proximity and expended effort independently enhanced participants' self-reported desire to obtain the reward, and their self-reported frustration and response vigor (key-press force) following blocking. In experiment 2, we used functional magnetic resonance imaging (fMRI) to show that both proximity and expended effort modulated brain responses to blocked reward in regions implicated in animal models of reactive aggression, including the amygdala, midbrain periaqueductal grey (PAG), insula and prefrontal cortex. Our findings suggest that frustration may serve an energizing function, translating unfulfilled motivation into aggressive-like surges via a cortical, amygdala and PAG network.

Using the moving window incorporated neural network to forecast the population behavior of Nostocales spp. in the River Darling, Australia

The paper demonstrates the nonstationarity of algal population behaviors by analyzing the historical populations of Nostocales spp. in the River Darling, Australia. Freshwater ecosystems are more likely to be nonstationary, instead of stationary. Nonstationarity implies that only the near past behaviors could forecast the near future for the system. However, nonstionarity was not considered seriously in previous research efforts for modeling and predicting algal population behaviors. Therefore the moving window technique was incorporated with radial basis function neural network (RBFNN) approach to deal with nonstationarity when modeling and forecasting the population behaviors of Nostocales spp. in the River Darling. The results showed that the RBFNN model could predict the timing and magnitude of algal blooms of Nostocales spp. with high accuracy. Moreover, a combined model based on individual RBFNN models was implemented, which showed superiority over the individual RBFNN models. Hence, the combined model was recommended for the modeling and forecasting of the phytoplankton populations, especially for the forecasting.

Modeling phytoplankton dynamics in the River Darling (Australia) using the radial basis function neural network

A radial basis function neural network was employed to model the abundance of cyanobacteria. The trained network could predict the populations of two bloom forming algal taxa with high accuracy, Nostocales spp. and Anabaena spp., in the River Darling, Australia. To elucidate the population dynamics for both Nostocales spp. and Anabaena spp., sensitivity analysis was performed with the following results. Total Kjeldahl nitrogen had a very strong influence on the abundance of the two algal taxa, electrical conductivity had a very strong negative relationship with the population of the two algal species, and flow was identified as one dominant factor influencing algal blooms after a scatter plot revealed that high flow could significantly reduce the algal biomass for both Nostocales spp. and Anabaena spp. Other variables such as turbidity, color, and pH were less important in determining the abundance and succession of the algal blooms.

Population genetic structure of Carchesium polypinum (Ciliophora : Peritrichia) in four chinese lakes inferred from ISSR fingerprinting: High diversity but low differentiation

Although the peritrichous ciliate Carchesium polypinum is common in freshwater, its population genetic structure is largely unknown. We used inter-simple sequence repeat (ISSR) fingerprinting to analyze the genetic structure of 48 different isolates of the species from four lakes in Wuhan, central China. Using eight polymorphic primers, 81 discernible DNA fragments were detected, among which 76 (93.83%) were polymorphic, indicating high genetic diversity at the isolate level. Further, Nei's gene diversity (h) and Shannon's Information index (I) between the different isolates both revealed a remarkable genetic diversity, higher than previously indicated by their morphology. At the same time, substantial gene flow was found. So the main factors responsible for the high level of diversity within populations are probably due to conjugation (sexual reproduction) and wide distribution of swarmers. Analysis of molecular variance (AMOVA) showed that there was low genetic differentiation among the four populations probably due to common ancestry and flooding events. The cluster analysis and principal component analysis (PCA) suggested that genotypes isolated from the same lake displayed a higher genetic similarity than those from different lakes. Both analyses separated C. polypinum isolates into subgroups according to the geographical locations. However, there is only a weak positive correlation between the genetic distance and geographical distance, suggesting a minor effect of geographical distance on the distribution of genetic diversity between populations of C. polypinum at the local level. In conclusion, our studies clearly demonstrated that a single morphospecies may harbor high levels of genetic diversity, and that the degree of resolution offered by morphology as a marker for measuring distribution patterns of genetically distinct entities is too low.

Regulation by hetC of genes required for heterocyst differentiation and cell division in Anabaena sp strain PCC 7120

Unlike those of the wild-type strain, proheterocysts of the Anabaena sp. strain PCC 7120 hetC strain keep dividing. ftsZ, the most critical cell division gene, is up-regulated in hetC proheterocysts. Heterocyst differentiation genes hglD, hglE, patB, nijB, and xisA are no longer expressed in the hetC mutant. hetC also regulates the expression of patA, a pattern formation gene.

Photoregulated or energy dependent process of hormogonia differentiation in Nostoc sphaeroides Kutzing (Cyanobacterium)

Hormogonium, which was thought to play an important role in the dispersal and survival of these microorganisms in their natural habitats, is a distinguishable developmental stage of heterocystous cyanobacteria. The present study examined the effects of different light conditions and sugars on the differentiation of Nostoc sphaeroides Kutzing to the hormogonia stage. Results showed that differentiation of hormogonia was light dependent in the absence of sugar, but that close to 100% of cyanobacteria differentiated to hormogonia in the presence of glucose or sucrose, irrespective of the light conditions. This differentiation was inhibited, even in the presence of sugars, upon application of an inhibitor of respiration. Following the testing of different sugars, the effects of different lights were examined. It was found that 5 10 μ mol.m(-2)• s(-1) photon flux density was optimal for hormogonia differentiation. One hundred percent differentiation was obtained with white light irradiation, in contrast with irradiation with green light (80% differentiation) and red light (0-10% differentiation). Although they showed different efficiencies in inducing hormogonia differentiation in N. sphaeroides, the green and red radiation did not display antagonistic effects. When the additional aspect of time dependence was investigated through the application of different light radiations and an inhibitor of protein synthesis, it was found that the initial 6 h of the differentiation process was crucial for hormogonia differentiation. Taken together, these results show that hormogonia differentiation in N. sphaeroides is either a photoregulated or an energy dependent process.

Microsatellites assessment of Chinese sturgeon (Acipenser sinensis Gray) genetic variability

Four microsatellites were used to examine the genetic variability of the spawning stocks of Chinese sturgeon, Acipenser sinensis, from the Yangtze River sampled over a 3-year period (1999-2001). Within 60 individuals, a total of 28 alleles were detected over four polymorphic microsatellite loci. The number of alleles per locus ranged from 4 to 15, with an average allele number of 7. The number of genotypes per locus ranged from 6 to 41. The genetic diversity of four microsatellite loci varied from 0.34 to 0.67, with an average value of 0.54. For the four microsatellite loci, the deviation from the Hardy-Weinberg equilibrium was mainly due to null alleles. The mean number of alleles per locus and the mean heterozygosity were lower than the average values known for anadromous fishes. Fish were clustered according to their microsatellite characteristics using an unsupervised 'Artificial Neural Networks' method entitled 'Self-organizing Map'. The results revealed no significant genetic differentiation considering genetic distance among samples collected during different years. Lack of heterogeneity among different annual groups of spawning stocks was explained by the complex age structure (from 8 to 27 years for males and 12 to 35 years for females) of Chinese sturgeon, leading to formulate an hypothesis about the maintenance of genetic diversity and stability in long-lived animals.

Genetic differentiation in populations of the cestode Bothriocephalus acheilognathi (Cestoda, Pseudophyllidea) as revealed by eight microsatellite markers

The genetic structure of populations of the fish cestode, Bothriocephalus acheilognathi collected from Bailianhe Reservoir (BLH), Changshou (CSH) and Liangzi (LZH) Lakes was investigated by using 8 microsatellite loci. A total of 108 adult worms were genotyped at each of the 8 loci. For the 3 populations, the mean number of alleles per locus ranged from 2.38 to 5.5, and the mean expected heterozygosity ranged from 0.432 to 0.559. The average polymorphic information content (PIC) was from 0.384 to 0.492. The significant F-is values indicated non-random mating within LZH and BLH populations. On the other hand, when samples were further classified into subpopulations at the level of host fish species, no or little heterozygote deficiency was detected at most loci, showing that cross-fertilization, predominantly, but not exclusively, must have occurred within the subpopulations. Microsatellite markers also revealed an unexpected high level of genetic differentiation, as measured by R-st and N-m values or by deltau(2) genetic distance among subpopulations from different hosts. Factors influencing the population genetic structure and the parasite host specificity are discussed.

Long-term toxic impact of 2,3,7,8-tetrachlorodibenzo-p-dioxin on the reproduction, sexual differentiation, and development of different life stages of Gobiocypris rarus and Daphnia magna

The sexual ratio of Gobiocypris rarus exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and 17 beta -estradiol from embryo to sexually mature revealed feminization and overdevelopment of connective tissue in male fish gonad in 2-30 pg/L TCDD concentration range. Daphnia magna was not sensitive to the high dose of TCDD (0.1-1000 ng/ml), but the reproduction of D. magna treated with TCDD decreased after the 8th day. 7-Ethoxyresorufin-O-deethylase (EROD) activities in newly fertilized eggs of G. rarus exposed to TCDD dosage groups (1000-100,000 pg/L) were significantly induced and increased with TCDD concentrations at the early life stage, while no difference was found between low TCDD dosage groups (<100 pg/L), but a good relationship between the EROD activity and the TCDD concentration was observed during a long-term developmental stage. There was a pericardial edema formed in a 2-week yolk-sac at the concentration of 1000 pg/L TCDD. But in the exposure group (2 pg/L TCDD for 120 days), the cell nuclei of hepatocytes was far from the center and packed toward the cell membrane; the cristae of most mitochondria in the cell dropped and collapsed; the rough endoplasmic reticulum broke into fragments; and numerous lipid droplets formed in the cell. (C) 2001 Academic Press.