C3: Speech-Guided KV Retrieval for Long-Context SpeechLLM Inference
cs.LG, cs.AI, cs.SD, eess.AS
Submitted: 2026-10-30
Updated: 2026-10-30
Terminology
Sources
- The MSP-Podcast Corpus
- DeepSeek-V3 Technical Report
- UNIQUE: Universal Top-k Sparse Attention for Training-free Inference and Sparsity-aware Training
- Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs
- AudioMarathon: A Comprehensive Benchmark for Long-Context Audio Understanding and Efficiency in Audio LLMs
- Murmur: Efficient Long-Context ASR Inference with Hierarchical Chunking
- LycheeCluster: Efficient Long-Context Inference with Structure-Aware Chunking and Hierarchical KV Indexing
- CompressKV: Semantic Retrieval Heads Know What Tokens are Not Important Before Generation
- Voxtral
- KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache
- Compress the Cache, Not the Speech Embedding: KV Compression for Efficient Speech LLMs
- Long-Form Speech Generation with Spoken Language Models
- MLS: A Large-Scale Multilingual Dataset for Speech Research
- ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference
- Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference
- TRADE: Transducer-Augmented Decoder for Speech LLM
- Gemma 4 Technical Report
- AudioKV: KV Cache Eviction in Efficient Large Audio Language Models
- Qwen2.5-Omni Technical Report
- Qwen3-Omni Technical Report
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis