Unit Selection with Hierarchical Cascaded Long Short Term Memory Bidirectional Recurrent Neural Nets

Pollet, Vincent; Zovato, Enrico; Irhimeh, S.; Batzu, Pier Domenico

doi:10.21437/interspeech.2017-428

Cited by 10 publications

(6 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Innovative uses of Long short-term memory recurrent neural networks (LSTM-RNNs) have been explored to various speech applications as can be seen in Reference [21], where the proposed architecture can synthesise natural sounding speech without requiring utterance-level batch processing. Likewise, an LSTM's performance is enhanced via hierarchical cascading in Reference [22].…”

Section: Related Workmentioning

confidence: 99%

A Context-Aware Conversational Agent in the Rehabilitation Domain

et al. 2019

View full text Add to dashboard Cite

Conversational agents are reshaping our communication environment and have the potential to inform and persuade in new and effective ways. In this paper, we present the underlying technologies and the theoretical background behind a health-care platform dedicated to supporting medical stuff and individuals with movement disabilities and to providing advanced monitoring functionalities in hospital and home surroundings. The framework implements an intelligent combination of two research areas: (1) sensor- and camera-based monitoring to collect, analyse, and interpret people behaviour and (2) natural machine–human interaction through an apprehensive virtual assistant benefiting ailing patients. In addition, the framework serves as an important assistant to caregivers and clinical experts to obtain information about the patients in an intuitive manner. The proposed approach capitalises on latest breakthroughs in computer vision, sensor management, speech recognition, natural language processing, knowledge representation, dialogue management, semantic reasoning, and speech synthesis, combining medical expertise and patient history.

show abstract

Section: Related Workmentioning

confidence: 99%

A Context-Aware Conversational Agent in the Rehabilitation Domain

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Moreover, there have been various efforts towards speech generation and the most popular ones include HMM-based speech synthesis models [29], specialized incremental dialogue systems [12], and DNN-based systems [34]. State-of-theart approaches for text-to-speech leverage various forms of DNNs; in [20] a system is presented that builds models based on waveforms, in [28] the models are based on spectrograms, while in [23] hierarchical cascading is used to improve an LSTM network. Furthermore, day-to-day voice enabled products like Apple's Siri, Amazon's Alexa/Echo, Microsoft's Cortana and Google's Assistant are joined by more healthcare oriented agents like Aiva 2 , Merit 3 , Suki 4 and Robin 5 .…”

Section: Related Workmentioning

confidence: 99%

A smart dialogue-competent monitoring framework supporting people in rehabilitation

Mavropoulos

Meditskos

Kamateri

et al. 2019

Proceedings of the 12th ACM International Conference on PErvasive Technologies Related to Assistive Environments

View full text Add to dashboard Cite

In this paper, we present work in progress on the development of a smart monitoring framework to support people with motor disabilities and their caregivers in clinical and non-clinical rehabilitation and care environments. The innovation of the platforms lies in the combination of smart monitoring solutions, such as activity recognition and lifestyle tracking, with an intelligent virtual agent that aims to empower and motivate people in need through personalized feedback and responses, as well as to assist caregivers and clinicians to easily collect information about the patients. The proposed system exploits and combines state-ofthe-art technologies in speech recognition and synthesis, knowledge representation and reasoning, dialogue management and sensor data analysis, infusing clinical knowledge and patient history. Aiming for a practical, acceptable solution, the proposed system takes into account aspects of integration, security and privacy.

show abstract

“…Hybrid synthesis systems make use of the advantages of both SPSS and unit selection based speech synthesis. The hybrid TTS synthesis system applies the statistical acoustic model to generate either speech parameter trajectories [16,17], or unit embeddings [18][19][20][21] to guide the unit selection process. In this paper, we propose a hybrid TTS synthesis system which uses an improved sequence-to-sequence (textto-feature) acoustic model to generate the speech parameter trajectory and trajectory tiling based waveform concatenation to synthesize the final waveform.…”

Section: Introductionmentioning

confidence: 99%

A New High Quality Trajectory Tiling Based Hybrid TTS In Real Time

Xie

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

A trajectory tiling based, hybrid TTS is revisited in this study for improving its synthesis performance. A combination of Transformer encoder and RNN based decoder architecture where two-level, at both word and Chinese phonetic alphabet letter levels, linguistic representation is exploited to generate a cogent and smooth speech parameter trajectory. And then a segment candidate lattice is constructed by minimizing the log spectral distortion of mel-spectrograms and RMSE of F0 between the generated trajectory and candidates. Normalized cross-correlation is used to find the best sequence of "waveform tiles" in the lattice for synthesizing the final speech waveforms. Subjective A/B preference tests show that the new hybrid system outperforms our earlier trajectory-tiling hybrid baseline TTS (67% vs 11%) and the state-of-the-art, real-time TTS system constructed with Tacotron 2 and LPC-Net (56% vs 27%).

show abstract

Unit Selection with Hierarchical Cascaded Long Short Term Memory Bidirectional Recurrent Neural Nets

Cited by 10 publications

References 10 publications

A Context-Aware Conversational Agent in the Rehabilitation Domain

A Context-Aware Conversational Agent in the Rehabilitation Domain

A smart dialogue-competent monitoring framework supporting people in rehabilitation

A New High Quality Trajectory Tiling Based Hybrid TTS In Real Time

Contact Info

Product

Resources

About