期满 发表于 2025-3-23 11:35:49

Sustainable Design for Global Equilibriumdesign where each class is characterized by an exponential-family distribution discussed in Chapter 1. The next chapter extends the results here into the more difficult but practically more useful case of hidden Markov models (HMMs).

权宜之计 发表于 2025-3-23 14:04:37

http://reply.papertrans.cn/29/2813/281227/281227_12.png

碎石头 发表于 2025-3-23 20:06:47

http://reply.papertrans.cn/29/2813/281227/281227_13.png

无脊椎 发表于 2025-3-23 22:38:02

CSR, Sustainability, Ethics & GovernanceThe basic development of growth transformation (GT)-based optimization for the hidden Markov model (HMM) and of the unified objective function has been presented in earlier chapters, where several practical considerations and implementation issues are left to this chapter.

craving 发表于 2025-3-24 04:03:12

http://reply.papertrans.cn/29/2813/281227/281227_15.png

使腐烂 发表于 2025-3-24 09:24:12

Synthesis Lectures on Speech and Audio Processinghttp://image.papertrans.cn/e/image/281227.jpg

–FER 发表于 2025-3-24 12:47:11

http://reply.papertrans.cn/29/2813/281227/281227_17.png

vascular 发表于 2025-3-24 17:58:43

http://reply.papertrans.cn/29/2813/281227/281227_18.png

镇压 发表于 2025-3-24 21:55:08

https://doi.org/10.1007/978-3-030-94818-4HMMs). These are: maximum mutual information (MMI), minimum classification error (MCE), and minimum phone error/minimum word error (MPE/MWE). We also compare our unified form of these objective functions with another popular unified form in the literature.

燕麦 发表于 2025-3-25 00:52:13

http://reply.papertrans.cn/29/2813/281227/281227_20.png
页: 1 [2] 3 4
查看完整版本: Titlebook: Discriminative Learning for Speech Recognition; Theory and Practice Xiaodong He,Li Deng Book 2008 Springer Nature Switzerland AG 2008