This paper presents a novel discriminative training technique for noisy speech recognition. First, we define a Frame Margin Probability (FMP) which denotes the difference of score of a frame on its right model and on its competing model. The frames with negative FMP values are regarded as confusable frames and the frames with positive FMP values are regarded as discriminable frames. Second, the...

This report describes the implementation of a discriminative HMM parameter estimation technique known as Frame Discrimination (FD) for large vocabulary speech recognition, and reports improvements in accuracy over ML-trained and MMI-trained models. Features of the implementation include the use of an algorithm called the Roadmap algorithm which selects the most important Gaussians for a given i...

Multi-task learning (MTL) can be an effective way to improve the generalization performance of singly learning tasks if the tasks are related, especially when the amount of training data is small. Our previous work applied MTL to the joint training of triphone and trigrapheme acoustic models using deep neural networks (DNNs) for low-resource speech recognition. Significant recognition improveme...

Mobile users with single antennas can use spatial transmission diversity through cooperative space-time encoded transmission. In this paper, we present an end-to-end performance analysis of two-hop asynchronous cooperative diversity with regenerative relays over Rayleigh block-flat-fading channels, in which a precoding frame-based scheme with packet-wise encoding is used. This precoding is base...

H. Heckhausen and J. Kuhl's (1985) goal typology provided the conceptual foundation for this research, which examined the independent and integrated effects of achievement orientation and goal-setting approaches on trainees' self-regulatory activity. Using a complex computer-based simulation, the authors examined the effects of 3 training design factors--goal frame, goal content, and goal proxi...

This paper describes a method to incorporate the HMM output constraints in frame based hybrid NN/HMM systems during training. While usually the NN parameters are adjusted to maximize the cross-entropy between the frame target probabilities and the network predictions assuming statistically independent outputs in time, in the approach described here the full likelihood of the utterance(s) using ...

This paper describes the use of a low-dimensional vector representation of sentence acoustics to control the output of a feed-forward deep neural network text-to-speech system on a sentence-by-sentence basis. Vector representations for sentences in the training corpus are learned during network training along with other parameters of the model. Although the network is trained on a frame-by-fram...

Segments that span contiguous parts of inputs, such as phonemes in speech, named entities in sentences, actions in videos, occur frequently in sequence prediction problems. Segmental models, a class of models that explicitly hypothesizes segments, have allowed the exploration of rich segment features for sequence prediction. However, segmental models suffer from slow decoding, hampering the use...

We describe a new technique for improving statistical machine translation training by adopting scores from a recent crosslingual semantic frame based evaluation metric, XMEANT, as outside probabilities in expectation-maximization based ITG (inversion transduction grammars) alignment. Our new approach strongly biases early-stage SMT learning towards semantically valid alignments. Unlike previous...

A fast adaptive Tomlinson Harashima (T-H) precoder structure is presented for indoor wireless communications, where the channel may vary due to rotation and small movement of the mobile terminal. A frequency-selective slow fading channel which is time-invariant over a frame is assumed. In this adaptive T-H precoder, feedback coefficients are updated at the end of every uplink frame by using sys...