A5100319050- Speech and Audio Processing
- Blind Source Separation Techniques
- Music and Audio Processing
- Text and Document Classification Technologies
- Neural Networks and Applications
- Speech Recognition and Synthesis
- Neural dynamics and brain function
- Topic Modeling
- Neuroscience and Music Perception
- Natural Language Processing Techniques
- EEG and Brain-Computer Interfaces
New York University
2018-2024
Japan Advanced Institute of Science and Technology
2022-2024
University of Tsukuba
2024
When we vocalize, our brain distinguishes self-generated sounds from external ones. A corollary discharge signal supports this function in animals; however, humans, its exact origin and temporal dynamics remain unknown. We report electrocorticographic recordings neurosurgical patients a connectivity analysis framework based on Granger causality that reveals major neural communications. find reproducible source for across multiple speech production paradigms localized to the ventral motor...
The superior temporal gyrus (STG) region of cortex critically contributes to speech recognition. In this work, we show that a proposed WaveNet, with limited available data, is able reconstruct stimuli from STG intracranial recordings. We further investigate the impulse response fitted model for each recording electrode and observe phoneme level temporospectral tuning properties recorded area cortex. This discovery consistent previous studies implicating posterior (pSTG) in phonetic...
Fake audio detection (FAD) aims to detect fake speech generated by advanced voice conversion and text-to-speech technologies. Recently, the quality of synthesized has significantly improved due remarkable development deep neural networks. However, it is still easy for humans identify perceiving pathological prosody in a voice. Pathological related amplitude frequency perturbation (AFP) provides essential cues speech. This paper proposed analyze AFP differences using jitter shimmer features....
Named Entity Recognition and Relation Extraction are two crucial challenging subtasks in Information Extraction. Despite the successes achieved by traditional approaches, fundamental research questions remain open. First, most recent studies use parameter sharing for a single subtask or shared features both subtasks, ignoring their semantic differences. Second, information interaction mainly focuses on leaving fine-grained among subtask-specific of encoding subjects, relations, objects...
The superior temporal gyrus (STG) region of cortex critically contributes to speech recognition. In this work, we show that a proposed WaveNet, with limited available data, is able reconstruct stimuli from STG intracranial recordings. We further investigate the impulse response fitted model for each recording electrode and observe phoneme level temporospectral tuning properties recorded area cortex. This discovery consistent previous studies implicating posterior (pSTG) in phonetic...