arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2207.05784cs.SDcs.LGeess.AS
收藏

面向低资源设备的基于二值神经网络的蒸馏非语义语音嵌入

Distilled Non-Semantic Speech Embeddings with Binary Neural Networks for Low-Resource Devices

University of California Los Angeles(加利福尼亚大学洛杉矶分校);Eindhoven University of Technology(埃因霍温理工大学)

Harlin Lee, Aaqib Saeed

AI总结:

本文提出基于二值神经网络的 BRILLsson 模型,通过知识蒸馏从大规模 TRILLsson 模型学习非语义语音表示,以仅 2MB 大小和低于 8ms 延迟在多个基准任务上达到与大规模模型相当的性能。

摘要:

本文介绍了 BRILLsson,一种新颖的基于二值神经网络的表示学习模型,适用于广泛的非语义语音任务。我们利用知识蒸馏,从大规模实值 TRILLsson 模型中训练该模型,且仅使用了训练 TRILLsson 所用数据集的一小部分。由此得到的 BRILLsson 模型大小仅为 2MB,延迟低于 8ms,使其适合部署在可穿戴设备等低资源设备中。我们在八个基准任务上评估了 BRILLsson(包括但不限于口语语言识别、情感识别、健康状况诊断和关键词识别),并证明我们提出的超轻量、低延迟模型能够达到与大规模模型相当的性能。

英文摘要:

This work introduces BRILLsson, a novel binary neural network-based representation learning model for a broad range of non-semantic speech tasks. We train the model with knowledge distillation from a large and real-valued TRILLsson model with only a fraction of the dataset used to train TRILLsson. The resulting BRILLsson models are only 2MB in size with a latency less than 8ms, making them suitable for deployment in low-resource devices such as wearables. We evaluate BRILLsson on eight benchmark tasks (including but not limited to spoken language identification, emotion recognition, health condition diagnosis, and keyword spotting), and demonstrate that our proposed ultra-light and low-latency models perform as well as large-scale models.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑