Time delay neural network (TDNN) has been proven to be efficient in learning discriminative speaker embeddings. One of its successful variant, ECAPA-TDNN, achieved state-of-the-art performance on speaker verification tasks at the cost of much higher computational complexity and slower inference speed. This makes it inadequate for scenarios with demanding inference rate and limited computational resources. We are thus interested in finding an architecture that can achieve the performance of ECAPA-TDNN and the efficiency of vanilla TDNN. In this paper, we propose an efficient network based on context-aware masking, namely CAM++, which uses densely connected time delay neural network (D-TDNN) as backbone and adopts a novel multi-granularity pooling to capture contextual information at different levels. Extensive experiments on two public benchmarks, VoxCeleb and CN-Celeb, demonstrate that the proposed architecture outperforms other mainstream speaker verification systems with lower computational cost and faster inference speed.

翻译：时延神经网络（TDNN）已被证明能有效学习判别性说话人嵌入。其成功变体之一ECAPA-TDNN在说话人验证任务上取得了最优性能，但代价是更高的计算复杂度和更慢的推理速度，这使其难以适用于推理速率要求高、计算资源受限的场景。因此，我们致力于寻找一种既能达到ECAPA-TDNN性能又具备原始TDNN效率的架构。本文提出一种基于上下文感知掩码的高效网络——CAM++，该网络以密集连接时延神经网络（D-TDNN）为主干，并采用新型多粒度池化机制捕获不同层级的上下文信息。在VoxCeleb和CN-Celeb两个公开基准上的大量实验表明，本文提出的架构以更低的计算成本和更快的推理速度超越了其他主流说话人验证系统。