16 papers
FlowGRPO for Vocal Separation: From Reward Analysis to Reward Design
Lishi Zuo, Chong-Xin Gan, Zezhong Jin +3
The Flow Matching objective supervises the velocity field at intermediate states, whereas waveform metrics are evaluated after decoding and may be non-differentiable. To align Flow…
TS-SP: Learning Speaker-Preserving Representations in Audio Large Language Models
Junjie Li, Zheng Liang, Zhe Li +2
Audio large language models (ALLMs) can understand speech content, yet their ability to use speaker identity for verification remains limited. We propose TS-SP (Two-Stage Speaker P…
A Unified Uncertainty-Aware Back-End for Speaker Verification: Scoring, Normalization, and Calibration
Junjie Li, Kong Aik Lee
Speaker verification back-ends commonly combine similarity scoring, score normalization, and calibration. However, speaker embeddings extracted from real-world utterances have tria…
Decoupled Latent Flow Matching for Few-Step Joint Vocal-Accompaniment Separation
Lishi Zuo, Youzhi Tu, Lu Yi +4
Generative modeling provides a flexible way to model mixture-conditioned source distributions, but iterative diffusion and flow matching models are costly for long music signals. T…
Learning to Prefer Reliably: Error-Augmented Emotion Preference Optimization with Calibrated Fusion
Zilong Huang, Junyi Peng, Junjie Li +5
Emotion preference learning uses pairwise comparisons between candidate descriptions to align multimodal large language models (MLLMs) with human judgments of open-ended emotion de…
Beyond Residual Connections: Manifold-Constrained Hyper-Connections for Robust Speaker Representation Learning
Zezhong Jin, Xiaoyu Wang, Zhe Li +4
Residual connections are fundamental to deep speaker recogni- tion models, such as ECAPA-TDNN and ResNet. However, standard identity mapping limits information flow to a sin- gle p…
EmoEUS: Uncertainty Supervision for Multimodal Emotion Recognition in Conversation
Zilong Huang, Kong Aik Lee, Junjie Li +2
Multimodal emotion recognition in conversation (MERC) can leverage multimodal and contextual cues to boost recognition performance. However, existing fusion approaches in MERC ofte…
EII-SCL: Harnessing Emotional Inertia for Multimodal Emotion Recognition in Conversation
Zilong Huang, Kong Aik Lee, Chong-Xin Gan +3
Multimodal emotion recognition in conversation (MERC) achieves accurate predictions by integrating multimodal and contextual information in dialogues. While current MERC approaches…
Text-Independent Speaker Verification Using Discrete Audio Tokens
Zheng Liang, Junjie Li, Kong Aik Lee
Neural audio codecs (NACs) enable efficient audio compression and have achieved success in downstream tasks such as speech synthesis. However, their discrete representations consis…
Towards Robust Uncertainty-Aware Speaker Modeling
Junjie Li, Yang Xiao, Kong Aik Lee
Speaker embeddings aggregate frame-level acoustic features into compact representations for speaker recognition. Recent uncertainty-aware speaker modeling approaches further charac…
UNet-Based Fusion and Exponential Moving Average Adaptation for Noise-Robust Speaker Recognition
Chong-Xin Gan, Peter Bell, Man-Wai Mak +4
The joint training of speech enhancement and speaker embedding networks for speaker recognition is widely adopted under noisy acoustic environments. While effective, this paradigm…
Xi+: Uncertainty Supervision for Robust Speaker Embedding
Junjie Li, Kong Aik Lee, Duc-Tuan Truong +2
There are various factors that can influence the performance of speaker recognition systems, such as emotion, language and other speaker-related or context-related variations. Sinc…
Subband Architecture Aided Selective Fixed-Filter Active Noise Control
Hong-Cheng Liang, Man-Wai Mak, Kong Aik Lee
The feedforward selective fixed-filter method selects the most suitable pre-trained control filter based on the spectral features of the detected reference signal, effectively avoi…
Bayesian Learning for Domain-Invariant Speaker Verification and Anti-Spoofing
Jin Li, Man-Wai Mak, Johan Rohdin +2
The performance of automatic speaker verification (ASV) and anti-spoofing drops seriously under real-world domain mismatch conditions. The relaxed instance frequency-wise normaliza…
On the effectiveness of enrollment speech augmentation for Target Speaker Extraction
Junjie Li, Ke Zhang, Shuai Wang +3
Deep learning technologies have significantly advanced the performance of target speaker extraction (TSE) tasks. To enhance the generalization and robustness of these algorithms wh…
VoxGenesis: Unsupervised Discovery of Latent Speaker Manifold for Speech Synthesis
Weiwei Lin, Chenhang He, Man-Wai Mak +2
Achieving nuanced and accurate emulation of human voice has been a longstanding goal in artificial intelligence. Although significant progress has been made in recent years, the ma…