Knowledge Commons of Institute of Automation,CAS
Lead ASR Models to Generalize Better Using Approximated Bias-Variance Tradeof | |
Wang FY(王方圆)1![]() ![]() | |
2023 | |
会议名称 | ICONIP 2023 |
会议日期 | 2023.11.13 |
会议地点 | changsha,China |
会议举办国 | 日本 |
摘要 | The conventional recipe for Automatic Speech Recognition (ASR) models is to 1) train multiple checkpoints on a training set while relying on a validation set to prevent over fitting using early stopping and 2) average several last checkpoints or that of the lowest validation losses to obtain the final model. In this paper, we rethink and update the early stopping and checkpoint averaging from the perspective of the bias-variance tradeoff. Theoretically, the bias and variance represent the fitness and variability of a model and the tradeoff of them determines the overall generalization error. But, it’s impractical to evaluate them precisely. As an alternative, we take the training loss and validation loss as proxies of bias and variance and guide the early stopping and checkpoint averaging using their tradeoff, namely an Approximated Bias-Variance Tradeoff ApproBiVT). When evaluating with advanced ASR models, our recipe provides 2.5%–3.7% and 3.1%–4.6% CER reduction on the AISHELL-1 and AISHELL-2, respectively (The code and sampled unaugmented training sets used in this paper will be public available on GitHub). |
七大方向——子方向分类 | 语音识别与合成 |
国重实验室规划方向分类 | 语音语言处理 |
是否有论文关联数据集需要存交 | 否 |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/57379 |
专题 | 复杂系统认知与决策实验室_听觉模型与认知计算 |
作者单位 | 1.中国科学院自动化研究所 2.广播科学院互联网所 |
第一作者单位 | 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Wang FY,Ming Hao,Yuhai Shi,et al. Lead ASR Models to Generalize Better Using Approximated Bias-Variance Tradeof[C],2023. |
条目包含的文件 | 下载所有文件 | |||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
iconip2023_published(1933KB) | 会议论文 | 开放获取 | CC BY-NC-SA | 浏览 下载 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论