Back in 2010, we noted that Baidu, the Chinese search and digital commerce giant, was staffing up to make speech recognition and natural language understanding core to its service offerings. Five years into its efforts, the introduction of an intelligent assistant called Duer shows that its technology is ready for prime time and that the people who are downloading its mobile app are the initial beneficiaries. Duer is bundled into the latest versions of the Baidu search app for smartphones. It represents an elegant marriage of the work that Baidu has done in speech recognition with significant enhancements to its search services that support the “parsing” of spoken input on the front end and and matching it with indexed tags, terms or topics derived from Web-based sites and content. That’s how Duer will help mobile searchers determine what hotels are “pet friendly” or what restaurants are best suited for vegans.

As reported in this article in AdAge.com, Baidu is the “third largest digital advertising player” on the planet, behind only Google and Facebook. These three companies are willingly offering products and service that condition all all of us to be comfortable using the most natural means to search, find and carry out conversations that culminate in transactions. Apple’s Siri is over seven years old and steadily extending its reach and ability to commercial conversations. By saying “OK Google” the general public is discovering that it can ask questions, give orders, dictate messages and carry out commerce reliably. As noted here last week, Facebook encourages avid users of Facebook Messenger to text their questions, suggestions and instructions to virtual assistants who can help them answer questions, refine searches and complete transactions. Duer serves the same functions and, collectively, all these offerings are refining, replacing or simply augmenting more conventional search engines.

Baidu envisions a world where Web-based content (and everything else) is totally “indexed.” The results are illustrated in this video, which leverages natural language understanding, machine translation, location awareness, navigation and “deep learning” and then ups the anti a strong dose of augmented reality, heads up displays in automobiles and indoor location in retail stores. Like Google, Baidu is willingly investing the billions of dollars it generates from digital marketing and advertising into “intelligent assistance” technologies.

Duer’s introduction coincides with a major product announcement from Apple that will make Siri a core part of the the natural interface for Apple TV. As Gadget Lab’s Tim Moynihan reported in Wired, Apple’s intent is to add a natural front end to the home entertainment console. With interactive video encroaching on messaging services and digital commerce, the addition of a Siri-like interface is a natural. But why stop with “Siri-like” when Apple can offer the original. The demo addressed control of video content but, as Amazon demonstrated with “Alexa” on its Echo home appliance, introduction of a speech-enabled natural interface to a home electronics device quickly lends itself to experimentation which, in turn, becomes the ideal source material for understanding input and constantly improving responses.

Success will breed success. The more Duers, Alexas, Siris and Echos we encounter “in the wild” of our own homes, cars and smartphones the better results we will accomplish. They are the universal and ubiquitous front-ends for intelligent assistance.

Categories: Intelligent Assistants, Articles

Tags: Baidu, Duer, Natural Language Understanding, Siri