Abstract
We present a compact, quantization-ready acoustic scene classification (ASC) framework that couples an efficient student network with a learned teacher ensemble and knowledge distillation. The student backbone uses stacked depthwise-separable "expand-depthwise-project" blocks with global response normalization to stabilize training and improve robustness to device and noise variability, while a global pooling head yields class logits for efficient edge inference. To inject richer inductive bias, we assemble a diverse set of teacher models and learn two complementary fusion heads: z1, which predicts per-teacher mixture weights using a student-style backbone, and z2, a lightweight MLP that performs per-class logit fusion. The student is distilled from the ensemble via temperature-scaled soft targets combined with hard labels, enabling it to approximate the ensemble's decision geometry with a single compact model. Evaluated on the TAU Urban Acoustic Scenes 2022 Mobile benchmark, our approach achieves state-of-the-art (SOTA) results on the TAU dataset under matched edge-deployment constraints, demonstrating strong performance and practicality for mobile ASC.
Abstract (translated)
我们提出了一种紧凑且适合量化的声学场景分类（ASC）框架，该框架结合了一个高效的“学生”网络与一个学习到的“教师”模型集合以及知识蒸馏技术。学生的主体采用堆叠的深度可分离"扩展-深度卷积-投影"(expand-depthwise-project)块，并使用全局响应规范化来稳定训练并提高对设备和噪声变化的鲁棒性，同时一个全局池化头部产生用于高效边缘推理的类别得分(logits)。为了注入更丰富的归纳偏置，我们组装了一组多样化的教师模型，并学习两个互补的融合头部：z1预测每个教师混合权重（使用学生样式的主体），而z2是一个轻量级MLP，执行每类logits融合。通过温度缩放后的软目标结合硬标签，学生从集合中蒸馏出来，使其能够用单个紧凑模型逼近集合的决策几何形状。 在TAU城市声学场景2022移动基准测试上评估我们的方法，在符合边缘部署约束条件下，该方法在TAU数据集上实现了最先进的（SOTA）结果，展示了在移动ASC中的强大性能和实用性。
URL
https://arxiv.org/abs/2512.13905
