Psychoacoustically Aligned Latent Smoothing for Adversarial Robustness of Full-Duplex Speech-to-Speech Dialogue Models
cs.SD, cs.AI, cs.CL, cs.HC
Submitted: 2026-09-23
Updated: 2026-09-23
Terminology
Sources
- Moshi: a speech-text foundation model for real-time dialogue
- Voice Jailbreak Attacks Against GPT-4o
- Universal and Transferable Adversarial Attacks on Aligned Language Models
- Qwen2-Audio Technical Report
- ANCER: Anisotropic Certification via Sample-wise Volume Maximization
- Certified Defense via Latent Space Randomized Smoothing with Orthogonal Encoders
- GLM-4-Voice: Towards Intelligent and Human-Like End-to-End Spoken Chatbot
- Full-Duplex-Bench: A Benchmark to Evaluate Full-duplex Spoken Dialogue Models on Turn-taking Capabilities
- Seamless Interaction: Dyadic Audiovisual Motion Modeling and Large-Scale Dataset
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- MUSAN: A Music, Speech, and Noise Corpus
Related papers
- Few-Shot Open-Set Audio Classification via Transductive Prototype Refinement and Class Logit Enhancement
- Spectral Masking and Interpolation Attack (SMIA): A Black-box Adversarial Attack against Voice Authentication and Anti-Spoofing Systems
- AVMeme Exam: A Multimodal Multilingual Multicultural Benchmark for LLMs' Contextual and Cultural Knowledge and Thinking
- SoundWeaver: Compositional Warm-Starting for Text-to-Audio Diffusion Serving
- WASIL: In-the-Wild Arabic Spoken Interactions with LLMs
- Efficient Audiovisual Speech Processing via MUTUD: Multimodal Training and Unimodal Deployment