ZET-Speech: Zero-shot adaptive Emotion-controllable Text-to-Speech Synthesis with Diffusion and Style-based Models

INTERSPEECH 2023 ZET-Speech: Zero-shot adaptive Emotion-controllable Text-to-Speech Synthesis with Diffusion and Style-based Models Minki Kang, Wooseok Han, Sung Ju Hwang, Eunho Yang   Emotional Text-To-Speech (TTS) is an important task in the development of systems (e.g., human-...

SGEM: Test-Time Adaptation for Automatic Speech Recognition via Sequential-Level Generalized Entropy Minimization

INTERSPEECH 2023 SGEM: Test-Time Adaptation for Automatic Speech Recognition via Sequential-Level Generalized Entropy Minimization  Changhun Kim, Joonhyung Park, Hajin Shim and Eunho Yang   Automatic speech recognition (ASR) models are frequently exposed to data di...

Learning Missing Modal Electronic Health Records with Unified Multi-modal Data Embedding and Modality-Aware Attention

MLHC 2023 Learning Missing Modal Electronic Health Records with Unified Multi-modal Data Embedding and Modality-Aware Attention Kwanhyung Lee, Soojeong Lee, Sangchul Hahn, Heejung Hyun, Edward Choi, Byungeun Ahn, Joohyung Lee   Electronic Health Record (EHR) provides abundant inf...

Traversing Between Modes in Function Space for Fast Ensembling

ICML 2023  (TML4H workshop)  Traversing Between Modes in Function Space for Fast Ensembling  Eunggu Yun, Hyungi Lee, Giung Nam, Juho Lee    Deep ensemble is a simple yet powerful way to improve the performance of deep neural networks. Under this moti...

RGE: A Repulsive Graph Rectification for Node Classification via Influence

ICML 2023 RGE: A Repulsive Graph Rectification for Node Classification via Influence  Jaeyun Song*, Sung-Yub Kim* and Eunho Yang (*: equal contribution)   In real-world graphs, noisy connections are inevitable, which makes it difficult to obtain unbiased node repre...

Probabilistic Imputation for Time-series Classification with Missing Data

ICML 2023 Probabilistic Imputation for Time-series Classification with Missing Data  SeungHyun Kim, Hyunsu Kim, Eunggu Yun, Hwangrae Lee, Jaehun Lee, Juho Lee    Multivariate time series data for real-world applications typically contain a significant amount o...

Fighting Fire with Fire: Contrastive Debiasing without Bias-free Data via Generative Bias-transformation

ICML 2023 Fighting Fire with Fire: Contrastive Debiasing without Bias-free Data via Generative Bias-transformation  Yeonsung Jung, Hajin Shim, June Yong Yang and Eunho Yang   Deep neural networks (DNNs), despite their impressive ability to generalize over-capacity ...

Diffusion Video Autoencoders: Toward Temporally Consistent Face Video Editing via Disentangled Video Encoding

CVPR 2023 Diffusion Video Autoencoders: Toward Temporally Consistent Face Video Editing via Disentangled Video Encoding   Gyeongman Kim, Hajin Shim, Hyunsu Kim, Yunjey Choi, Junho Kim and Eunho Yang   Inspired by the impressive performance of recent face image...

WeavSpeech: Data Augmentation Strategy For Automatic Speech Recognition Via Semantic-Aware Weaving

ICASSP 2023 WeavSpeech: Data Augmentation Strategy For Automatic Speech Recognition Via Semantic-Aware Weaving Kyusung Seo, Joonhyung Park, Jaeyun Song and Eunho Yang   A cut-and-paste type of data augmentation strategy has attracted considerable attention in the vision...

Grad-StyleSpeech: Any-Speaker Adaptive Text-to-Speech Synthesis with Diffusion Models

ICASSP 2023 Grad-StyleSpeech: Any-Speaker Adaptive Text-to-Speech Synthesis with Diffusion Models Minki Kang, Dongchan Min, Sung Ju Hwang  There has been a significant progress in Text-To-Speech (TTS) synthesis technology in recent years, thanks to the advancement in ne...