Report of the baseline survey undertaken on the Chinese cage site in Napoleon Gulf, Northern Lake Victoria, 12 March 2012

The government of the People's Republic of China through a 2007 agreement with the Government of the Republic of Uganda, has establishment of an Agricultural Technology Demonstration Center (ATDC). The first phase covering the building of aquaculture infrastructure at Kajjansi ARDC is complete and the second operation phase has started in which facilities for cage culture have been set up in the Napoleon gulf, northern Lake Victoria near Jinja. The cage facility is aimed at boosting fish farming within the lake as a diversification to the traditional pond fish culture technology. NaFIRRI scientists as well as Chinese experts undertook a baseline survey in the chosen cage site on 12 March 2012. The survey covered determination of water depth, water transparency, measurement of selected physical-chemical parameters (temperature,dissolved oxygen, conductivity and pH; determination of the nutrient status and study of algae, invertebrate and fish communities at the site. Materials and methodologies used in the survey were based on the Standard Operating Procedures (SOPs) of NaFIRRI. The study area was divided into three study sites. Site 1 (upstream) was at 8.9 metre depth while site 2 (proposed cage site) and site 3 (downstream) were 6 and 4.3 metres deep respectively. Water transparency was lowest at site 1 (1.58 m) and highest at site 3 (1.64 m). Dissolved oxygen at the three sites ranged from 6.0 to 8 mg/I. Water temperature profiles fluctuated within narrow limits between 26.5 and 27.5 DC. Measurements of pH were between 7 (neutral) and 8 (alkaline) while electrical conductivity was between 98 and 101 uS/em. These observed physical-chemical parameters at the study site were considered suitable for cage fish rearing purposes. Nitrite-nitrogen levels varied within narrow limits from 0.043 to 0.0453 mgtl. Similarly, Ammonia-nitrogen varied between 0.015 and 0.0185 mg/1. Soluble reactive phosphorus (SRP) level was highest at site 3 (O.012mgll) compared to that at sites 1 and 2 (0.009mgll). Total suspended solids (TSS) were higher at site 1 (83.3mgll), thereafter decreasing to lower levels at sites 2 (24.8mgtl) and 3 (19.8mgl) respectively. The nutrient level results observed here all fall below the maximum permissible limits by NEMA and therefore the site is recommended for cage culture The algal community was constituted by four major groups: Blue greens,Greens, Cryptophytes, and Diatoms with blue greens as the common and dominant group. High algal biomass (19944961 ugtL) of the dominant blue green algae was observed at site 1 compared site 2 and 3 (58655.2 & 27487. 7 ugtL) respectively. Occurrence of toxicin producing algae: microsytis and cylindrospermopsis in the proposed cage area was considered to be of not much significance as their concentrations were below harmful levels. However, monitoring their presence, biomass and seasonality will be critical in order to follow when and where they occur and at what time of the year for ease of management of the cages

Preliminary estimates of cetacean abundance in the northern Gulf of Mexico, and of selected cetacean species in the U.S. Atlantic Exclusive Economic Zone from vessel surveys

The Southeast Fisheries Science Center (SEFSC) initiated annual, vessel-based visual sampling surveys of northern Gulf of Mexico marine mammals in 1990 and conducted a similar survey in U.S. Atlantic Exclusive Economic Zone (EEZ) waters from Miami, Florida, to Cape Hatteras, North Carolina, in 1992. The primary goal of these surveys was to meet Marine Mammal Protection Act requirements for estimating abundance and monitoring trends of marine mammal stocks in United States waters. The surveys were designed to collect: 1) marine mammal sighting data to estimate abundance and to determine distribution and diversity; and 2) environmental data to evaluate factors which may affect the distribution, abundance and diversity of marine mammals. The preliminary analyses for abundance estimation from the 1990-1993 surveys are presented in this report.

Estimates of cetacean abundance in the northern Gulf of Mexico from vessel surveys

The Southeast Fisheries Science Center (SEFSC) initiated annual, vessel-based visual sampling surveys of northern Gulf of Mexico marine mammals in 1990. The primary goal of these surveys was to meet Marine Mammal Protection Act requirements for estimating abundance and monitoring trends of marine mammal stocks in United States waters. The surveys were designed to collect: 1) marine mammal sighting data to estimate abundance and to determine distribution and diversity; and 2) environmental data to evaluate factors which may affect the distribution, abundance and diversity of marine mammals. The analyses for abundance estimation from the 1991-1994 surveys are presented in this report.

Statistical parametric speech synthesis based on speaker and language factorization

An increasingly common scenario in building speech synthesis and recognition systems is training on inhomogeneous data. This paper proposes a new framework for estimating hidden Markov models on data containing both multiple speakers and multiple languages. The proposed framework, speaker and language factorization, attempts to factorize speaker-/language-specific characteristics in the data and then model them using separate transforms. Language-specific factors in the data are represented by transforms based on cluster mean interpolation with cluster-dependent decision trees. Acoustic variations caused by speaker characteristics are handled by transforms based on constrained maximum-likelihood linear regression. Experimental results on statistical parametric speech synthesis show that the proposed framework enables data from multiple speakers in different languages to be used to: train a synthesis system; synthesize speech in a language using speaker characteristics estimated in a different language; and adapt to a new language. © 2012 IEEE.

Phrase-based statistical language generation using graphical models and active learning

Most previous work on trainable language generation has focused on two paradigms: (a) using a statistical model to rank a set of generated utterances, or (b) using statistics to inform the generation decision process. Both approaches rely on the existence of a handcrafted generator, which limits their scalability to new domains. This paper presents BAGEL, a statistical language generator which uses dynamic Bayesian networks to learn from semantically-aligned data produced by 42 untrained annotators. A human evaluation shows that BAGEL can generate natural and informative utterances from unseen inputs in the information presentation domain. Additionally, generation performance on sparse datasets is improved significantly by using certainty-based active learning, yielding ratings close to the human gold standard with a fraction of the data. © 2010 Association for Computational Linguistics.

Improving LVCSR system combination using neural network language model cross adaptation

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple subsystems developed at different sites. Cross system adaptation can be used as an alternative to direct hypothesis level combination schemes such as ROVER. The standard approach involves only cross adapting acoustic models. To fully exploit the complimentary features among sub-systems, language model (LM) cross adaptation techniques can be used. Previous research on multi-level n-gram LM cross adaptation is extended to further include the cross adaptation of neural network LMs in this paper. Using this improved LM cross adaptation framework, significant error rate gains of 4.0%-7.1% relative were obtained over acoustic model only cross adaptation when combining a range of Chinese LVCSR sub-systems used in the 2010 and 2011 DARPA GALE evaluations. Copyright © 2011 ISCA.

Use of contexts in language model interpolation and adaptation

Language models (LMs) are often constructed by building multiple individual component models that are combined using context independent interpolation weights. By tuning these weights, using either perplexity or discriminative approaches, it is possible to adapt LMs to a particular task. This paper investigates the use of context dependent weighting in both interpolation and test-time adaptation of language models. Depending on the previous word contexts, a discrete history weighting function is used to adjust the contribution from each component model. As this dramatically increases the number of parameters to estimate, robust weight estimation schemes are required. Several approaches are described in this paper. The first approach is based on MAP estimation where interpolation weights of lower order contexts are used as smoothing priors. The second approach uses training data to ensure robust estimation of LM interpolation weights. This can also serve as a smoothing prior for MAP adaptation. A normalized perplexity metric is proposed to handle the bias of the standard perplexity criterion to corpus size. A range of schemes to combine weight information obtained from training data and test data hypotheses are also proposed to improve robustness during context dependent LM adaptation. In addition, a minimum Bayes' risk (MBR) based discriminative training scheme is also proposed. An efficient weighted finite state transducer (WFST) decoding algorithm for context dependent interpolation is also presented. The proposed technique was evaluated using a state-of-the-art Mandarin Chinese broadcast speech transcription task. Character error rate (CER) reductions up to 7.3 relative were obtained as well as consistent perplexity improvements. © 2012 Elsevier Ltd. All rights reserved.

Syllable language models for Mandarin speech recognition: exploiting character language models.

Mandarin Chinese is based on characters which are syllabic in nature and morphological in meaning. All spoken languages have syllabiotactic rules which govern the construction of syllables and their allowed sequences. These constraints are not as restrictive as those learned from word sequences, but they can provide additional useful linguistic information. Hence, it is possible to improve speech recognition performance by appropriately combining these two types of constraints. For the Chinese language considered in this paper, character level language models (LMs) can be used as a first level approximation to allowed syllable sequences. To test this idea, word and character level n-gram LMs were trained on 2.8 billion words (equivalent to 4.3 billion characters) of texts from a wide collection of text sources. Both hypothesis and model based combination techniques were investigated to combine word and character level LMs. Significant character error rate reductions up to 7.3% relative were obtained on a state-of-the-art Mandarin Chinese broadcast audio recognition task using an adapted history dependent multi-level LM that performs a log-linearly combination of character and word level LMs. This supports the hypothesis that character or syllable sequence models are useful for improving Mandarin speech recognition performance.