Deep neural networks (DNNs) have been playing a significant role in acoustic modeling. Convolutional neural networks (CNNs) are the advanced version of DNNs that achieve 4–12% relative gain in the word error rate (WER) over DNNs. Existence of spectral variations and local correlations in speech signal makes CNNs more capable of speech recognition. Recently, it has been demonstrated that bidirectional long short-term memory (BLSTM) produces higher recognition rate in acoustic modeling because they are adequate to reinforce higher-level representations of acoustic (...) data. Spatial and temporal properties of the speech signal are essential for high recognition rate, so the concept of combining two different networks came into mind. In this paper, a hybrid architecture of CNN-BLSTM is proposed to appropriately use these properties and to improve the continuous speech recognition task. Further, we explore different methods like weight sharing, the appropriate number of hidden units, and ideal pooling strategy for CNN to achieve a high recognition rate. Specifically, the focus is also on how many BLSTM layers are effective. This paper also attempts to overcome another shortcoming of CNN, i.e. speaker-adapted features, which are not possible to be directly modeled in CNN. Next, various non-linearities with or without dropout are analyzed for speech tasks. Experiments indicate that proposed hybrid architecture with speaker-adapted features and maxout non-linearity with dropout idea shows 5.8% and 10% relative decrease in WER over the CNN and DNN systems, respectively. (shrink)

Business leadership increasingly requires a master’s degree in business and graduate management admission test scores continue to be an important component of applications for admission to such programs. Given the ubiquitous use of GMAT scores as gatekeepers for business leadership, GMAT scores are likely to influence organizational ethical behavior through gender, cultural, and other biases in the GMAT. There is little prior literature in this area and we contribute by empirically documenting that GMAT scores are negatively related to the cultural (...) dimensions of masculinity and power distance and are positively related to math literacy, uncertainty avoidance, and individualism. We estimate that cultural factors may account for as much as an 80-point difference in cross-national mean GMAT scores which are also related negatively to local language literacy, national educational spending, wealth per capita, wealth inequality, and gender development. Most interestingly, we also find a significant negative association of GMAT scores with ethical orientation. These findings have important implications for business schools and corporate ethics and leadership. (shrink)

This paper implements the continuous Hindi Automatic Speech Recognition (ASR) system using the proposed integrated features vector with Recurrent Neural Network (RNN) based Language Modeling (LM). The proposed system also implements the speaker adaptation using Maximum-Likelihood Linear Regression (MLLR) and Constrained Maximum likelihood Linear Regression (C-MLLR). This system is discriminatively trained by Maximum Mutual Information (MMI) and Minimum Phone Error (MPE) techniques with 256 Gaussian mixture per Hidden Markov Model(HMM) state. The training of the baseline system has been done using (...) a phonetically rich Hindi dataset. The results show that discriminative training enhances the baseline system performance by up to 3%. Further improvement of ~7% has been recorded by applying RNN LM. The proposed Hindi ASR system shows significant performance improvement over other current state-of-the-art techniques. (shrink)

Particular visions of urban development are often codified in multi-year resource management policies. These policies, and the negotiations leading to them, are based in specific problem frames and narratives with long legacies. As conditions change and knowledge improves, there is often a need to revisit how problems, opportunities, and development pathways were defined historically, and to consider the viability of alternative pathways for development. In this article, we examine the case of agriculture near Metropolitan Phoenix, in the Central Arizona region, (...) to highlight how frames and narratives embedded in policy can reinforce particular development pathways, even as information, conditions, and values evolve. Using expert interviews and secondary data, we document alternative frames and narratives that may offer different pathways for development and sustainability in the region. By highlighting alternative narratives, we demonstrate the uncertainties and limitations associated with all narratives about development pathways, and explore the possibilities that narrative shifts can alter future outcomes. (shrink)

In this paper, a HIV-TB co-infection model is explored which incorporates a non-linear treatment rate for TB. We begin with presenting a HIV-TB co-infection model and analyze both HIV and TB sub-models separately. The basic reproduction numbers corresponding to HIV-only, TB-only and the HIV-TB full model are computed. The disease-free equilibrium point of the HIV sub-model is shown to be locally as well as globally asymptotically stable when its corresponding reproduction number is less than unity. The HIV-only model exhibits a (...) transcritical bifurcation. On the other hand, for the TB sub-model, the disease-free equilibrium point is locally asymptotically stable but may not be globally asymptotically stable. We have also analyzed the full HIV-TB co-infection model. Numerical simulations are performed to investigate the effect of treatment rate in the presence of resource limitation for TB infected individuals, which emphasize the fact that to reduce co-infection from the population programs to accelerate the treatment of TB should be implemented. (shrink)