Reward-Tilted On-Policy Distillation for Acoustic Grounding in Audio-Language Models

arXiv:2609.28778 · cs.SD, cs.CL · Submitted 2026-09-23 · Read on arXiv

cs.SD, cs.CL

Submitted: 2026-09-23

Updated: 2026-09-23

Code: https://github.com/KaiyangLi1992/RT-OPD

Terminology

• Distillation — 34× in this paper · explained in 2 episodes
• On-Policy Distillation — 10× in this paper · explained in 3 episodes
• Accuracy — 8× in this paper · explained in 1 episode
• Benchmark — 8× in this paper · explained in 2 episodes
• Grounding — 5× in this paper · explained in 3 episodes
• Reward-Tilted On-Policy Distillation — 4× in this paper
• Ablation — 2× in this paper · explained in 1 episode
• Alignment — 2× in this paper · explained in 1 episode
• GRPO — 2× in this paper · explained in 6 episodes
• Hallucination — 2× in this paper · explained in 3 episodes
• Knowledge Distillation — 2× in this paper · explained in 11 episodes
• Reinforcement Learning — 2× in this paper · explained in 4 episodes
• Ablation study — 1× in this paper · explained in 1 episode
• Adaptation — 1× in this paper · explained in 1 episode
• Core — 1× in this paper · explained in 2 episodes
• Exposure Bias — 1× in this paper · explained in 2 episodes
• Multimodal Reasoning — 1× in this paper · explained in 5 episodes

Sources

• Qwen2-Audio Technical Report
• Qwen2.5-Omni Technical Report
• Summary of DCASE 2026 Task 5: Audio-Dependent Question Answering
• Step-Audio 2 Technical Report
• MiMo-Audio: Audio Language Models are Few-Shot Learners
• Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models
• Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
• Kimi-Audio Technical Report
• CORD: Bridging the Audio-Text Reasoning Gap via Weighted On-policy Cross-modal Distillation
• X$^3$-OPD: Distilling Reasoning into Large Audio-Language Models via On-Policy Alignment
• CAAD: Contrastive Audio-Aware Distillation for Efficient Speech Language Models
• Visual-Advantage On-Policy Distillation for Vision-Language Models
• Learning beyond Teacher: Generalized On-Policy Distillation with Reward Extrapolation

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