At the beginning of 2017, Chinese tech company Baidu, the largest provider of Chinese language internet search as well as other digital products and services, committed to emerging business sectors such as artificial intelligence (AI) and machine learning. Since China has 731 million internet users, almost twice the U.S. population, Baidu's data set is capable of fueling AI algorithms to make them even better. With this focus on artificial intelligence, Baidu is exploring some very intriguing applications for artificial intelligence and machine learning including in their offices where facial recognition technology makes standard ID cards unnecessary and allows you to order tea from a vending machine.
They have also recruited top AI talent including one of the world's most notable AI pioneers Lu Qi, who was previously a Microsoft executive before he became Baidu's COO in January 2017. Qi will step down in July 2018 for personal reasons. Although he was only at Baidu for a short time, he helped chart a clear strategy for the company's AI operations that will continue. Here are a few ways Baidu uses artificial intelligence and machine learning.
DuerOS is Baidu's voice assistant
Since Baidu can leverage its expansive data set, its voice assistant called DuerOS has accumulated more conversation-based skill sets than Alexa, Siri or Cortana. Partnering with other tech companies is one way Baidu hopes to accelerate innovation. They have teamed up with more than 130 DuerOS partners, and the voice assistant is in more than 100 brands of appliances such as refrigerators, TVs, and speakers. Since homes in India, Japan, Europe, and Brazil are more like homes in China, there may be better opportunities for DuerOS to globalize since Alexa, Cortana and Echo are optimized for American households. At CES 2018, Baidu debuted its DuerOS-powered smart screen called Little Fish VS1. This technology can recognize and respond to individual faces.
Mobile partners to accelerate AI-powered devices
Unlike its competitors, Baidu was steadfast in its commitment to desktops and missed the shift to mobile. To survive, Baidu needed a new strategy and artificial intelligence technology provided just the platform to turn the business around. That's one of the reasons Baidu has committed so aggressively to AI investment. Today, AI products and services are priorities to make them the core of the company's future. Now, they are partnering with Huawei to develop an open mobile AI platform to support the development of AI-powered smartphones and Qualcomm to optimize its DuerOS for IoT devices and smartphones using Qualcomm's Snapdragon Mobile Platform.
Even though automated driving is currently against the law in China, Baidu is working on autonomous-driving technology. Through the program called Apollo, Baidu's artificial intelligence technologies are made available to car makers for free as a brain for their cars. In exchange, Baidu gets access to the data to make their algorithms smarter. It is hoped that Apollo will give any car manufacturer a fair shot at creating a viable product, just like Android did for smartphone makers. Even with this accelerated approach, it is expected that fully autonomous cars won't be in production until 2020-2021.
Deep learning in the medical field
To find tumors and early warning signs for cancer, pathologists look at biopsied tissues magnified to 40X, but they are searching for micrometastases that can be as small as 1000 pixels in diameter. Deep learning-based algorithms can now be used to process the images and locate possible micrometastases. Baidu has developed a deep learning solution to predict if the cells are normal or cancerous. The results of using the algorithms in this way are promising-a lower number of false positives, better tumor localization scores, and improved efficiency. Computer vision that results in better detection of disease (or at least equal to human medical professionals) is a very compelling form of artificial intelligence, and one Baidu is actively working on.
Baidu recently spun-off its financial services arm on the heels of selling off other business units such as its food-delivery service Waimai and IQiyi Inc., a streaming video service comparable to Netflix. These moves are intended to streamline operations so Baidu can invest aggressively in artificial intelligence initiatives such as autonomous vehicles and voice assistants. By leveraging partnerships, its unique space in Chinese culture that is less averse to AI, the Chinese government's support as well as an enormous population that creates the data that allows AI insights, Baidu's bet on AI might just be the ticket to its long-term success.