Batched Speech Decisions Without Decoding: Single-Token Supervision Lets a Frozen LLM Hear Beyond the Transcript

cs.AI

Submitted: 2026-10-02

Updated: 2026-10-02

Comments: 5 pages, 2 figures, 3 tables. Submitted to ICASSP 2027. Code and weights: https://github.com/adventists-ai/duplexjev

Code: https://github.com/adventists-ai/duplexjev

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Sources

• Moshi: a speech-text foundation model for real-time dialogue
• Easy Turn: Integrating Acoustic and Linguistic Modalities for Robust Turn-Taking in Full-Duplex Spoken Dialogue Systems
• Qwen3-ASR Technical Report
• Distilling the Knowledge in a Neural Network
• Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models
• Qwen3 Technical Report
• VoxCPM: Tokenizer-Free TTS for Context-Aware Speech Generation and True-to-Life Voice Cloning

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems