Recently, Baidu has made breakthroughs in speech recognition technology, successfully "cross-border" image recognition technology into the field of speech, using deep convolutional neural network (Deep CNN) for acoustic modeling of speech recognition, based on its length and duration. The combination of memory unit (LSTM) and connected timing classification (CTC) end-to-end speech recognition technology, the error rate is reduced by 10%, greatly improving the performance of speech recognition products, is another significant after end-to-end speech recognition. Technological breakthroughs.
Deep CNN speech recognition modeling process
In recent years, the image recognition results using CNN technology are quite abundant. The increasingly deep CNN constantly refreshes the accuracy of image recognition. Taking face recognition as an example, the recognition accuracy is as high as 99.7%. However, the progress of CNN has not been fully applied in speech recognition. As an artificial intelligence company with in-depth research in voice technology, Baidu regards Deep CNN as the next breakthrough in speech recognition technology.
In the ImageNet competition, more and more CNN is constantly refreshing its performance.
In the end-to-end speech recognition technology in the commercial field, Baidu first tried to introduce a deeper CNN neural network, which reduced the error rate by 10%. The end-to-end technology uses a separate learning algorithm to complete all the processes from the task input to the output, reducing the intermediate unit and human intervention, and the model's effect is improved with the support of massive data. At present, Baidu's end-to-end technology is at the leading level in the industry. It is worth mentioning that the speech recognition is completed based on the speech spectrum after time-frequency analysis. The time spectrum analyzed by the whole speech signal is regarded as an image, and the widely used CNN in the image can be used to identify and overcome The problem of voice signal diversity, and the introduction of deeper CNN, the speech recognition performance has been significantly improved, as Dr. Li Xiangang, head of the recognition technology of Baidu Voice Technology Department, said: 'The Deeper, The Better'.
Different from academic research, Baidu's speech research and development is based on the practical application of technology, and the technical difficulty and degree of realization are higher. For speech recognition products, it is necessary to have a performance improvement on a large-scale voice database and a model suitable for the operation of voice online identification products. Baidu used thousands of hours of experimental research and verified in nearly 100,000 hours of product voice database, and sufficient voice data resources, making the speech recognition system based on end-to-end technology significantly better than the previous framework performance.
Baidu speech recognition technology annual iterative algorithm model
In addition, Baidu voice technology has significant advantages in data, computing power, and algorithms. Baidu has about 100,000 hours of precision-labeled voice data and a high-performance computing platform based on hundreds of GPUs. In terms of algorithms, Baidu is constantly optimizing and iterating model algorithms every year, and the speech recognition effect is significantly improved, leading the industry.
Previously, Baidu facilitated the use of end-to-end technology to develop Deep Speech 2 deep speech recognition technology to improve the accuracy of speech recognition in noisy environments. In a noisy environment, the error rate is lower than that of Google, Microsoft, and Apple's voice system. At present, Baidu's speech recognition accuracy rate is as high as 97%, and it has been listed as one of the top ten breakthrough technologies in 2016 by the American authoritative technology magazine "MIT Report". According to Dr. Li Xianang, the development of Deep Speech 3 is indeed being intensified, and the release of Deep CNN will not be excluded as a core component of Deep Speech 3.
In addition to technological breakthroughs, Baidu also actively promotes the popularity of voice interactions among users. Mobile phones such as Baidu, Baidu Input, Baidu Map, and Mi Mi have supported voice input functions, and this “cross-border†Deep CNN believes that it will soon Applied to Baidu products with a large user base.
Whether you have an Apple Ibook, an Apple Powerbook or indeed an Apple Macbook Air, we will have the Apple Laptop Charger suitable for your Apple laptop. Apple is a high quality brand, and therefore Apple laptop charger must be high quality too.
Apple laptop charger include apple macbook pro charger series and apple macbook air charger series. Yidashun not only can offer 45W 60W 85W old mac charger with magsafe 1.0 and 2.0 tip, but also can offer new 29W 30W 61W 87W USB C power Adapter. And also we can silkprint your logo on the chargers, and also can customize the color package.
If you want to look for a factory which can produce range models of the macbook charger, and also with high quality, contact Yidashun , we can offer you all kinds of replacement apple adapter. and also support you high quality with 2 years' warranty.
Yidahsun's laptop adapter is with smart IC to protect your laptop with over current protection, over load protection, short circuit protection, over heat protection. All our mac laptop charger is Brand New Replacement Product, works as Genuine parts, 100% OEM Compatible!
Apple Laptop Charger,Apple Computer Charger,Apple Macbook Charger,Mac Laptop Charger
Shenzhen Yidashun Technology Co., Ltd. , https://www.ydsadapter.com