CAJ | 학술논문

提出了一种最大后验(maximum a posteriori, MAP)估计和加权近邻回归(weighted neighbors regression, WNR)相结合的说话人自适应方法。在MAP自适应中，只有自适应数据对应的模型参数可以得到调整。针对这一缺点，提出一种基于变换的模型插值/平滑方法-WNR，利用模型近邻信息和MAP自适应结果，建立距离加权的回归模型，对没有自适应数据的模型完成模型调整。实验证明，该方法可以有效地提高MAP自适应的速度。在自适应数据为10句时，音节误识率降低近15%；而在自适应数据为250句时，误识率降低50%以上。此外，证明了向量域平滑(vector field smoothing, VFS)是WNR方法的一种退化的特例。
제출료일충최대후험(maximum a posteriori, MAP)고계화가권근린회귀(weighted neighbors regression, WNR)상결합적설화인자괄응방법。재MAP자괄응중，지유자괄응수거대응적모형삼수가이득도조정。침대저일결점，제출일충기우변환적모형삽치/평활방법-WNR，이용모형근린신식화MAP자괄응결과，건립거리가권적회귀모형，대몰유자괄응수거적모형완성모형조정。실험증명，해방법가이유효지제고MAP자괄응적속도。재자괄응수거위10구시，음절오식솔강저근15%； 이재자괄응수거위250구시，오식솔강저50%이상。차외，증명료향량역평활(vector field smoothing, VFS)시WNR방법적일충퇴화적특례。
This paper describes a novel speaker adaptation framework that combines the maximum a posteriori (MAP) estimation and wighted neighbor regression (WNR) methods. A great deal of adaptation data is required in MAP adaptation because only the parameters of those models with adaptation data can be updated. To alleviate this disadvantage, a technique called WNR is presented in which the parameter relationships between the speaker independent models and the speaker adaptation models are trained by applying distance weighted regression to a set of neighbor model parameters with and without MAP adaptation. The Chinese syllable recognition error is reduced nearly 15 percent with 10 adaptation utterances and more than 50 percent with 250 utterances. In addition, vector field smoothing (VFS) can be proved to be a degenerate case of WNR.