The task of personalized keyword detection system which also performs text dependent speaker verification (TDSV) has received substantial interest recently. Conventional approaches to this task involve the development of the TDSV and wakeup-word detection systems separately. In this paper, we show that TDSV and keyword spotting (KWS) can be jointly modeled using the convolutional long short term memory (CLSTM) model architecture, where an initial convolutional feature map is further processed by a LSTM recurrent network. Given a small amount of training data for developing the CLSTM system, we show that the model provides accurate detection of the presence of the keyword in spoken utterance. For the TDSV task, the MTL model can be well regularized using the CLSTM training examples for personalized wake up task. The experiments are performed for KWS wake up detection and TDSV using the combined speech recordings from Wall Street Journal (WSJ) and LibriSpeech corpus. In these experiments with multiple keywords, we illustrate that the proposed approach of MTL significantly improves the performance of previously proposed neural network based text dependent SV systems. We also experimentally illustrate that the CLSTM model provides significant improvements over previously proposed keyword detection systems as well (average relative improvements of 30% over previous approaches).

Abstract. Production parallel systems are space-shared and hence employ batch queues in which the jobs submitted to the systems are made to wait before execution. Thus, jobs submitted to parallel batch systems incur queue waiting times in addition to the execution times. Prediction of these queue waiting times is important to provide overall estimates to the users and can also help metaschedulers make scheduling decisions. Analyses of the job traces of supercomputers reveal that about 56 to 99% of the jobs incur queue waiting times of less than an hour. Hence, identifying these quick starters or jobs with short queue waiting times is essential for overall improvement on queue waiting time predictions. Existing strategies provide high overestimates of upper bounds of queue waiting times rendering the bounds less useful for jobs with short queue waiting times. In this work, we have developed an integrated framework that uses the job characteristics, and states of the queue and processor occupancy to identify and predict quick starters, and use the existing strategies to predict jobs with long queue waiting times. Our experiments with different production supercomputer job traces show that our prediction strategies can lead to correct identification of up to 20 times more quick starters and provide tighter bounds for these jobs, and thus result in up to 64% higher overall prediction accuracy than existing methods.

No abstract

No abstract

In music signal processing, singing voice detection and music source separation are widely researched topics. Recent progress in deep neural network based source separation has advanced the state of the performance in the problem of vocal and instrument separation, while the problem of joint source activity detection and separation remains unexplored. In this paper, we propose an approach to perform source activity detection using the high-dimensional embedding generated by Deep Attractor Network (DANet) when trained for music source separation. By defining both tasks together, DANet is able to dynamically estimate the number of outputs depending on the active sources. We propose an Expectation-Maximization (EM) training paradigm for DANet which further improves the separation performance of the original DANet. Experiments show that our network achieves higher source separation and comparable source activity detection against a baseline system.