The World is Not Mono: Enabling Spatial Understanding in Large Audio-Language Models
cs.SD, cs.AI
Submitted: 2026-01-06
Updated: 2026-11-07
Comments: NeurIPS 2026 camera-ready version. 35 pages, 4 figures
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- SoundSpaces: Audio-Visual Navigation in 3D Environments
- Qwen2-Audio Technical Report
- Learning Spatially-Aware Language and Audio Embeddings
- Clotho: An Audio Captioning Dataset
- FSD50K: An Open Dataset of Human-Labeled Sound Events
- Soft Adaptive Policy Optimization
- Audio Flamingo Next: Next-Generation Open Audio-Language Models for Speech, Sound, and Music
- Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models
- LoRA: Low-Rank Adaptation of Large Language Models
- A Track-Wise Ensemble Event Independent Network for Polyphonic Sound Event Localization and Detection
- LibriTTS-R: A Restored Multi-Speaker Text-to-Speech Corpus
- Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilities
- The Song Describer Dataset: a Corpus of Audio Captions for Music-and-Language Evaluation
- ZeRO: Memory Optimizations Toward Training Trillion Parameter Models
- Pyroomacoustics: A Python package for audio room simulations and array processing algorithms
- ACCDOA: Activity-Coupled Cartesian Direction of Arrival Representation for Sound Event Localization and Detection
- Multi-ACCDOA: Localizing and Detecting Overlapping Sounds from the Same Class with Auxiliary Duplicating Permutation Invariant Training
- Towards Spatial Audio Understanding via Question Answering
- SALMONN: Towards Generic Hearing Abilities for Large Language Models
- HuggingFace's Transformers: State-of-the-art Natural Language Processing
Related papers
- Speech-DRAME: A Framework for Human-Aligned Benchmarks in Speech Role-Play
- MuLaCover: Flexible Cover Song Generation via Symbolic Tonal Control
- Optimising Neural Speech Codecs for 300bps Communication using Reinforcement Learning
- 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