LY Corporation R&D

Our papers have been accepted at INTRSPEECH 2026 (external link).

Online Predictive Coding for Dual-Mode Self-Supervised Speech Models
Keita Goto, Takashi Maekaku, Jin Sakuma, Jinchuan Tian (CMU), Yusuke Shinohara, and Shinji Watanabe (CMU)

Refining Pseudo-Audio Prompts with Speech-Text Alignment for Text-Only Domain Adaptation in LLM-Based ASR
Ryo Magoshi (Kyoto University), Takashi Maekaku, and Yusuke Shinohara

Bagpiper-TTS: Natural Language Guided Universal Speech Synthesis
Jinchuan Tian (CMU), Haoran Wang (CMU), Siddhant Arora (CMU), Takashi Maekaku, Keita Goto, Jin Sakuma, Yusuke Shinohara, Chao-Han Huck Yang (NVIDIA Research), and Shinji Watanabe (CMU)

Investigating Human-Model Discrepancies in Speech Quality Assessment via Acoustic and Prosodic Perturbations
Masato Takagi (Nagoya Institute of Technology), Masaya Kawamura, Reo Shimizu (Tohoku University), and Yuma Shirahata

PASQA: Pitch-Accent-Focused Speech Quality Assessment Model Trained on Synthetic Speech with Accent Errors
Masaya Kawamura, Yuma Shirahata, Kentaro Mitsui, and Reo Shimizu

ProLAP: Probabilistic Language-Audio Pre-Training
Toranosuke Manabe (Keio University), Yuchi Ishikawa, Hokuto Munakata, Yoshimitsu Aoki (Keio University), and Tatsuya Komatsu

Aligning MusicLLM with Emotion using Instruction Tuning and Feedback-Driven Alignment
Takuya Hasumi and Welly Naptali