Tomoko Matsui

ORCID: 0000-0003-3201-6106
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Speech Recognition and Synthesis
  • Speech and Audio Processing
  • Music and Audio Processing
  • Speech and dialogue systems
  • Natural Language Processing Techniques
  • Language, Metaphor, and Cognition
  • Child and Animal Learning Development
  • Language Development and Disorders
  • Language, Discourse, Communication Strategies
  • Neural Networks and Applications
  • Gaussian Processes and Bayesian Inference
  • Text and Document Classification Technologies
  • Topic Modeling
  • Human Mobility and Location-Based Analysis
  • Climate Change Policy and Economics
  • Time Series Analysis and Forecasting
  • Land Use and Ecosystem Services
  • Anomaly Detection Techniques and Applications
  • Soil Geostatistics and Mapping
  • Neurobiology of Language and Bilingualism
  • Statistical Methods and Inference
  • Educational Strategies and Epistemologies
  • Autism Spectrum Disorder Research
  • Blind Source Separation Techniques
  • Urban Heat Island Mitigation

Eli Lilly (Japan)
2022-2025

The Institute of Statistical Mathematics
2015-2024

Hamamatsu University School of Medicine
2024

Chuo University
2023

Tokyo Gakugei University
2012-2022

University of Electro-Communications
2019

National Institute for Environmental Studies
2015

University College London
1993-2015

Nagoya University
2013

Kyoto University
2004-2010

A VQ (vector quantization)-distortion-based speaker recognition method and discrete/continuous ergodic HMM (hidden Markov model)-based ones are compared, especially from the viewpoint of robustness against utterance variations. It is shown that a continuous far superior to discrete HMM. also information on transitions between different states ineffective for text-independent recognition. Therefore, identification rates using strongly correlated with total number mixtures irrespective states....

10.1109/icassp.1992.226096 article EN 1992-01-01

Recently, a series of studies demonstrated false belief understanding in young children through completely nonverbal measures. These have revealed that younger than 3 years age, who consistently fail the standard verbal test, can anticipate others' actions based on their attributed beliefs. The current study examined whether with autism spectrum disorder (ASD), are known to difficulties may also show such action anticipation test. We presented video stimuli an actor watching object being...

10.1017/s0954579410000106 article EN Development and Psychopathology 2010-04-28

Methods that create models to specify both speaker and phonetic information accurately by using only a small amount of training data for each are investigated. For text-dependent recognition method, in which arbitrary key texts prompted from the recognizer, speaker-specific phoneme necessary identify text recognize speaker. Two methods making discussed: phoneme-adaptation phoneme-independent model speaker-adaptation universal models. The authors also investigate supplementing these adding...

10.1109/icassp.1993.319321 article EN IEEE International Conference on Acoustics Speech and Signal Processing 1993-01-01

Gaussian Processes (GPs) are Bayesian nonparametric models that becoming more and popular for their superior capabilities to capture highly nonlinear data relationships in various tasks, such as dimensionality reduction, time series analysis, novelty detection, well classical regression classification tasks. In this paper, we investigate the feasibility applicability of GP music genre emotion estimation. These two main tasks information retrieval (MIR) field. So far, support vector machine...

10.1109/access.2014.2333095 article EN cc-by-nc-nd IEEE Access 2014-01-01

It has been repeatedly shown that when asked to identify a protagonist's false belief on the basis of his statement, English-speaking 3-year-olds dismiss statement and fail attribute him belief. In present studies, we tested 3-year-old Japanese children in similar task, using statements accompanied by grammaticalized particles speaker (un)certainty, as everyday utterances. The were directly compared with same-aged German children, whose native language does not have epistemic concepts....

10.1111/j.1467-7687.2008.00812.x article EN Developmental Science 2009-02-13

A hearer's perception of an utterance as sarcastic depends on integration the heard statement, discourse context, and prosody utterance, well evaluation incongruity among these aspects. The effect in sarcasm comprehension is evident everyday conversation, but little known about its underlying mechanism or neural substrates. To elucidate underpinnings auditory modality, we conducted a functional MRI experiment with 21 adult participants. participants were provided short vignette which child...

10.1016/j.neuropsychologia.2016.04.031 article EN cc-by Neuropsychologia 2016-05-06

Public transport, or mass transit, is believed to bring several benefits, including energy efficiency, air quality and an economy of scale with respect total public transport costs. An important factor in providing the cost-efficient timely service, ability do resource planning optimization based on up-to-date data passenger counts flows. In this work we give a novel approach counting number passengers bus WiFi signatures from any mobile devices carried onto vehicle. This provides traffic...

10.1109/itsc.2017.8317687 article EN 2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC) 2017-10-01

This paper compares a VQ (vector quantization)-distortion-based speaker recognition method and discrete/continuous ergodic HMM (hidden Markov model)-based ones, especially from the viewpoint of robustness against utterance variations. The authors show that continuous is as robust VQ-distortion when enough data available far superior to discrete HMM. They also information on transitions between different states ineffective for text-independent recognition. Therefore, rates using are strongly...

10.1109/89.294363 article EN IEEE Transactions on Speech and Audio Processing 1994-07-01

Real-time urban climate monitoring provides useful information that can be utilized to help management personnel monitor and adapt their precautionary measures extreme events, including heatwaves. Fortunately, recently created social media platforms, such as Twitter, furnish real-time high-resolution spatial may for condition estimation. The objective of this paper was utilize geotagged tweets (participatory sensing data) temperature analysis. We first detected related heat (heat-tweets)....

10.1109/access.2016.2516918 article EN cc-by-nc-nd IEEE Access 2016-01-01

One significant problem for spoken language systems is how to cope with users' out-of-domain (OOD) utterances which cannot be handled by the back-end application system.In this paper, we propose a novel OOD detection framework, makes use of classification confidence scores multiple topics and applies linear discriminant model perform in-domain verification.The verification trained using combination deleted interpolation data minimum-classification-error training, does not require actual...

10.1109/tasl.2006.876727 article EN IEEE Transactions on Audio Speech and Language Processing 2006-12-19

The authors describe a VQ (vector-quantization)-based text-independent speaker recognition method which is robust against utterance variations. Three techniques are introduced to cope with temporal and text-dependent spectral First, either an ergodic hidden Markov model or voiced/unvoiced decision used classify input speech into broad phonetic classes. Second, new distance measure, the distortion-intersection measure (DIM), for calculating distortion of compared speaker-independent...

10.1109/icassp.1991.150355 article EN 1991-01-01

We develop new algorithms for spatial field reconstruction, exceedance level estimation and classification in heterogeneous (mixed analog & digital sensors) Wireless Sensor Networks (WSNs). consider physical phenomena which are observed by a WSN, meaning that it consists partially of sparsely deployed high-quality sensors low-quality sensors. The transmit their (continuous) noisy observations to the Fusion Centre (FC), while first perform simple thresholding operation then binary values over...

10.1109/tsp.2015.2412917 article EN cc-by IEEE Transactions on Signal Processing 2015-03-13

Abstract Developmental research suggests that young children tend to value dominant individuals over subordinates. This research, however, has nearly exclusively been carried out in Western cultures, and cross-cultural among adults revealed cultural differences the valuing of dominance. In particular, it seems Japanese culture, relative many values dominance less. We conducted two experiments test whether this difference would be observed preschoolers. Experiment 1, preschoolers France Japan...

10.1163/15685373-12340058 article EN Journal of Cognition and Culture 2019-08-07

Significant improvement in progression-free survival (PFS; primary end point) was reported the phase 3 RELAY study with ramucirumab (RAM) plus erlotinib (ERL) versus placebo (PL) untreated EGFR-mutated NSCLC (hazard ratio [HR] = 0.59, 95% confidence interval [CI]: 0.46-0.76, p < 0.0001), including Japanese subset. We report updated PFS and final overall (OS) for Patients (no central nervous system metastases) were randomized 1:1 (stratification included EGFR leucine to arginine substitution...

10.1016/j.jtocrr.2025.100819 article EN cc-by-nc-nd JTO Clinical and Research Reports 2025-02-28
Coming Soon ...