Speaker Recognition In Noisy Conditions With Limited Training Data

In this paper we present a novel method for performing speaker recognition with very limited training data and in the presence of background noise. Similarity-based speaker recognition is considered so that speaker models can be created with limited training speech data. The proposed similarity is a form of cosine similarity used as a distance measure between speech feature vectors. Each speech frame is modelled using subband features, and into this framework, multicondition training and optimal feature selection are introduced, making the system capable of performing speaker recognition in the presence of realistic, time-varying noise, which is unknown during training. Speaker identi?cation experiments were carried out using the SPIDRE database. The performance of the proposed new system for noise compensation is compared to that of an oracle model; the speaker identi?cation accuracy for clean speech by the new system trained with limited training data is compared to that of a GMM trained with several minutes of speech. Both comparisons have demonstrated the effectiveness of the new model. Finally, experiments were carried out to test the new model for speaker identi?cation given limited training data and with differing levels and types of realistic background noise. The results have demonstrated the robustness of the new system.

Effect on prescribing of the limited list in a computerised group practice

The prescribing of drugs in the therapeutic classes that are affected by the government's limited list was investigated in a computerised group practice of just over 3,000 patients. Prescribable drugs in categories that are affected by the list were identified for two consecutive six month periods before and one six month period after the introduction of the list. A significant decrease in the prescribing of cough and cold remedies, vitamins, and antacids occurred after the list was introduced, whereas no change occurred in the prescribing of laxatives, benzodiazepines, or analgesics. The prescribing of iron and penicillin increased significantly after the list was introduced, whereas the use of H2 antagonists and non-steroidal anti-inflammatory drugs showed no significant change.

Kinematic Analysis and Dimensional Synthesis of Exechon Parallel Kinematic Machine for Large Volume Machining

A parallel kinematic machine (PKM) topology can only give its best performance when its geometrical parameters are optimized. In this paper, dimensional synthesis of a newly developed PKM is presented for the first time. An optimization method is developed with the objective to maximize both workspace volume and global dexterity of the PKM. Results show that the method can effectively identify design parameter changes under different weighted objectives. The PKM with optimized dimensions has a large workspace to footprint ratio and a large well-conditioned workspace, hence justifies its suitability for large volume machining.

Robust Multimodal Person Identification With Limited Training Data

This paper presents a novel method of audio-visual feature-level fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there are limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new multimodal feature representation and a modified cosine similarity are introduced to combine and compare bimodal features with limited training data, as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal dataset created from the SPIDRE speaker recognition database and AR face recognition database with variable noise corruption of speech and occlusion in the face images. The system's speaker identification performance on the SPIDRE database, and facial identification performance on the AR database, is comparable with the literature. Combining both modalities using the new method of multimodal fusion leads to significantly improved accuracy over the unimodal systems, even when both modalities have been corrupted. The new method also shows improved identification accuracy compared with the bimodal systems based on multicondition model training or missing-feature decoding alone.

How Much Multiuser Diversity is Required for Energy Limited Multiuser Systems?

Multiuser diversity (MUDiv) is one of the central concepts in multiuser (MU) systems. In particular, MUDiv allows for scheduling among users in order to eliminate the negative effects of unfavorable channel fading conditions of some users on the system performance. Scheduling, however, consumes energy (e.g., for making users' channel state information available to the scheduler). This extra usage of energy, which could potentially be used for data transmission, can be very wasteful, especially if the number of users is large. In this paper, we answer the question of how much MUDiv is required for energy limited MU systems. Focusing on uplink MU wireless systems, we develop MU scheduling algorithms which aim at maximizing the MUDiv gain. Toward this end, we introduce a new realistic energy model which accounts for scheduling energy and describes the distribution of the total energy between scheduling and data transmission stages. Using the fact that such energy distribution can be controlled by varying the number of active users, we optimize this number by either i) minimizing the overall system bit error rate (BER) for a fixed total energy of all users in the system or ii) minimizing the total energy of all users for fixed BER requirements. We find that for a fixed number of available users, the achievable MUDiv gain can be improved by activating only a subset of users. Using asymptotic analysis and numerical simulations, we show that our approach benefits from MUDiv gains higher than that achievable by generic greedy access algorithm, which is the optimal scheduling method for energy unlimited systems. © 2010 IEEE.

An Investigation towards colossal success of Larsen and Toubro Limited: Human resource management perspective

Larsen and Toubro (L&T) Limited is India’s largest construction conglomerate. L&T’s expertise is harnessed to execute high value projects that demand adherence to stringent timelines in a scenario where disparate disciplines of engineering are required to be coordinated on a critical path. However, no company can acquire such a feat without systematic management of its human resource. An investigation on the human resource management practices in orienting L&T’s success can help to identify some of the ethical human resource practices, especially in the context of Indian market. Accordingly, a well-designed employee satisfaction survey was conducted for assessment of the HRM practices being followed in L&T. Unlike other companies, L&T aims to meet the long-term needs of its employees rather than short-term needs. There were however few areas of concerns, such as yearly appraisal system and equality to treat the employees. It is postulated that the inequality to treat the male and female employees is primarily a typical stereotype due to the fact that construction is conventionally believed to be a male dominant activity. A periodic survey intended to provide 360° feedback system can help to avoid such irregularities. This study is thus expected to provide healthy practices of HRM to nurture the young talents of India. This may help them to evaluate their decisions by analyzing the complex relationship between HRM practices and output of an organization.

Non-enzymatic glycosylation of skin collagen in patients with type 1 (insulin-dependent) diabetes mellitus and limited joint mobility

Forearm skin biopsies were obtained from diabetic subjects with and without limited joint mobility, and from non-diabetic control subjects. Collagen purified from these samples was assayed for non-enzymatic glycosylation. The level in all diabetic patients was significantly greater than that in control subjects (p less than 0.001), but those diabetic patients with limited joint mobility had a level of collagen glycosylation similar to that in those with normal joints (15.3 +/- 1.3 and 16.5 +/- 1.3 nmol fructose/10 mg protein, respectively; mean +/- SEM). Glycosylation of collagen in the diabetic patients correlated with glycosylated haemoglobin measured at the time of skin biopsy (r = 0.60). These results do not support the hypothesis that non-enzymatic glycosylation of collagen, as reflected by the ketoamine link, plays an important role in the development of limited joint mobility in diabetes.