AS squared D: Accelerating On-Demand Audio Understanding on Mobile Devices
cs.SD, cs.AI
Submitted: 2026-09-29
Updated: 2026-09-29
Code: https://github.com/alibaba/MNN
Sources
- Sequoia: Scalable, Robust, and Hardware-aware Speculative Decoding
- Segmentwise Pruning in Audio-Language Models
- Streaming End-to-end Speech Recognition For Mobile Devices
- Event-Grounded Question Answering over Long Audio via Structured Retrieval
- Locality Matters for Training-Free Audio Token Compression in Audio-Language Models
- AHASD: Asynchronous Heterogeneous Architecture for LLM Adaptive Drafting Speculative Decoding on Mobile Devices
- WhisperKit: On-device Real-time ASR with Billion-Scale Transformers
- Self-Speculative Decoding for LLM-based ASR with CTC Encoder Drafts
- Qwen3-ASR Technical Report
- OmniMem: Perturbation-aware Memory Compression for Streaming Audio-Visual LLMs
- Lever: Speculative LLM Inference on Smartphones
- LLMCad: Fast and Scalable On-device Large Language Model Inference
- Qwen2.5-Omni Technical Report
- Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inference
- Learning to Draft: Adaptive Speculative Decoding with Reinforcement Learning
- Memoro: Using Large Language Models to Realize a Concise Interface for Real-Time Memory Augmentation
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
- Remix the Timbre: Diffusion-Based Style Transfer Across Polyphonic Stems
- 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