9 papers
Concept-Constrained Prompt Learning for Few-Shot CLIP Adaptation
Na Sang, Ding Ma, Rui Sang +1
Few-shot prompt learning is an effective strategy for adapting CLIP to downstream tasks, but class-only prompt optimization can overfit base-class supervision and weaken transfer t…
Geometry-Aware Localized Watermarking for Copyright Protection in Embedding-as-a-Service
Zhimin Chen, Xiaojie Liang, Wenbo Xu +2
Embedding-as-a-Service (EaaS) has become an important semantic infrastructure for natural language and multimedia applications, but it is highly vulnerable to model stealing and co…
Membership Inference Attack Against Music Diffusion Models via Generative Manifold Perturbation
Yuxuan Liu, Peihong Zhang, Rui Sang +4
Membership inference attacks (MIAs) test whether a specific audio clip was used to train a model, making them a key tool for auditing generative music models for copyright complian…
TopSeg: A Multi-Scale Topological Framework for Data-Efficient Heart Sound Segmentation
Peihong Zhang, Zhixin Li, Yuxuan Liu +4
Deep learning approaches for heart-sound (PCG) segmentation built on time-frequency features can be accurate but often rely on large expert-labeled datasets, limiting robustness an…
DDSC: Dynamic Dual-Signal Curriculum for Data-Efficient Acoustic Scene Classification under Domain Shift
Peihong Zhang, Yuxuan Liu, Rui Sang +4
Acoustic scene classification (ASC) suffers from device-induced domain shift, especially when labels are limited. Prior work focuses on curriculum-based training schedules that str…
SceneGuard: Training-Time Voice Protection with Scene-Consistent Audible Background Noise
Rui Sang, Yuxuan Liu
Voice cloning technology poses significant privacy threats by enabling unauthorized speech synthesis from limited audio samples. Existing defenses based on imperceptible adversaria…