Baidu’s input method AI features new upgrade: unlimited dialects, more free input

In today’s world where mobile phones have evolved to be smart, what are the features we use most on our phones? The answer is the input method. As the medium that conveys our thoughts to mobile phones and other mobile phone users, mobile input methods play a vital role in the efficiency and experience of using them. So what should a good input method look like for the user? The answer is for input methods to become like humans who understand our minds. Can any input method really do that? Baidu’s input method, with its new AI features, may be putting this standard into practice.

Baidu’s input method is now in front of us with new AI features. Baidu’s input method has been upgraded to provide more convenient input experience, with three new dialects for Hebei, Kunming and Wuhan in the “Dialect Free Speech” module, as well as improved recognition accuracy for Fujian, Guangdong, Hunan, Jiangxi, Sichuan and Wu dialects. Existence.

In addition to support for multiple dialects, if you open the voice input feature now, you will also find new options for Chinese-English free speech, Japanese voice input, as well as Chinese-English, Chinese-Japanese, Chinese-Korean and English-to-Chinese and Chinese-Japanese to Chinese. This means that we can also use two foreign languages for voice input, and can switch between input and display text in multiple languages at will in real time, enabling a one-step process from voice translation to input.

This AI feature upgrade for Baidu’s input method is a quantum leap for the input method. As mentioned earlier, a good input method should be able to understand our minds like a human. When two people communicate with each other, it is possible to eliminate the influence of dialects, foreign languages and other distracting factors to a certain extent, without affecting the fluency of communication at all. Baidu input method through the AI upgrade to let the dialect free speech support nine dialects as well as many regions of heavy accent Mandarin, precisely so that the input method can exclude these interfering factors, the correct understanding of what we as users want to convey.

Baidu input method has been working to improve the input experience with AI, and has previously introduced an intelligent prediction feature that can predict what we want to say in our popular chat software based on the context before and after, which has garnered quite a bit of praise. And now voice input enhancements are leveraging AI to make the input method understand us better. Whether it is keyboard or voice input, there are corresponding AI technology in Baidu input method to enhance the experience, in fact, more than that, Baidu input method to create a variety of input experience including keyboard input, voice input, handwriting input, expression input (AR expression), action input (overhead handwriting) and other input methods, and in early 2019, Baidu input method announced to fully open the “AI The “full-sensory input 2.0” era, with the support of AI technology to make the input method more and more intelligent, these are through Baidu input method on a functional module upgrade to let us experience.

Why can Baidu input method do all this? Because Baidu input method for artificial intelligence layout early enough, and did a deep enough development. Looking back to 2019, Baidu had already released stream truncated multi-layer attention modeling, which improved the relative accuracy of online voice by 15%, surpassing its peers. And with the support of this technology, the accuracy of offline voice recognition reached the same level as online voice, and even if users are in environments like subways, elevators, underground tunnels, and other environments with poor networks, Baidu’s input method can still provide voice input services with an accuracy of more than 98%. Back to the present, we will also find that the upgraded Baidu input method can achieve offline Chinese and English free speech, and the accuracy is still quite high, this is the technical accumulation of Baidu input method.

If we divide input methods into different stages according to the experience of using them, then the first stage is the handwriting and keystroke input and the second stage is the voice input. The first stage is highly accurate but inefficient, while the second stage is highly efficient and the biggest problem is how to accurately convey the content in an input environment with thousands of people and thousands of languages. Baidu input method of AI technology for these two stages have different strengths, so bring the input method is a comprehensive upgrade.

Nowadays, Baidu input method can not only know what we want to say through the intelligent prediction function, but also can freely input various dialects through the dialect free speech function module, no longer need to manually switch special options, just leave everything to AI technology. Baidu input method is not only a simple input tool, but also supports Chinese, Cantonese, Sichuan, Northeast, Henan, Shandong, Xi’an, Hefei, and English, Korean, Japanese and Chinese translation, helping us to bridge the language gap and extend the concept of “input method”.

With the upgrade of AI technology, the new Baidu input method has brought an enhanced version of AI full-sensory input, which is no longer the traditional input method, but can make the mobile phone a more important communication tool and make the communication between people more smooth. In this era of information technology, Baidu input method is making information communication more and more efficient, as well as making people closer to each other.