With the development of network technology, the artificial intelligence (AI) technology is also undergoing constant updates and iterations, and bringing dramatic changes to social production and human life. Following the previous Unlock Secrets | Qing AI Open Voice Platform of PATEO CONNECT+ (Part I) that gives you a closer look at the PATEO Qing AI open voice platform and its attributes, this article will give you a detailed introduction of its configuration, operation process and technical strength.
Qing AI voice platform with a full process configuration
The PATEO Qing AI platform controls the whole flow of speech processing based on a full link configuration, composed of ASR (Automatic Speech Recognition), NLU (Natural Language Understanding) & skill services, TTS (Text to Speech) and front-end display, with all settings supporting online real-time (OLRT) updates and capable of immediately taking effect, so that the normal service operation in the production environment is not affected.
For ASR, one can set the NLU supplier, the duration of VAD (Voice Activity Detection), and the dialect, among others, corresponding to specified scenarios, use the recommended presets of the platform, or configure settings according to client’s requirements for each project.
For NLU and skill services, one can set the NLU supplier for a project and the range of skills covered in it. Based on the NLU-empowered basic semantic understanding, PATEO has independently developed its own skill services, as well as Central Console Module (CCM) and context-based dialog management module, with the aim of achieving optimal results in natural language dialogue scenarios.
For TTS, one can select supplier, speaker, dialect and so on as needed based on the recommended configuration presets of the platform.
How to operate the Qing AI open voice platform?
Creation in five steps & generation of exclusive AI SDK in five minutes
The Qing AI open voice platform provides developers and vehicle manufacturers with a platform for development of full voice capabilities, offers a range of services covering product introduction, functional development, experience testing, upgrade & iteration, product version management, data analysis, etc., and enables full-link customization, zero-cost development, as well as rapid implementation of in-vehicle voice solution.
Based on the platform, an exclusive in-vehicle intelligent voice product can be created in five steps, with capabilities open for independent selection and configuration; also, in as fast as five minutes, an exclusive AI SDK can be generated for local deployment, then the real interaction effects of the vehicle-borne AI-empowered product can be experienced in a real vehicle.
• Step 1: Complete configuration of the underlying capabilities of the Qing AI product — PATEO has been conducting deep cooperation with basic voice technology suppliers such as AISpeech, iFLYTEK and Cerence, to provide multiple offline engine & online recognition combination solutions;
• Step 2: Complete basic configuration for personalized AI — Allow different carmakers to personalize Wake-Up-Word (WUW), AI character (including virtual character), emotional TTS, speaker, etc., and open custom interface for them to upload their own AI & TTS voices;
• Step 3: Complete selection and configuration of skills for AI product — The Qing AI voice product covers scenarios in various fields like leisure & entertainment, business travel, and communication in life, supports continuous dialogue & multi-round interaction in all scenarios and closed-loop online payment in airport, hotel & other eco scenarios, thus creating full-scenario AI-empowered smart cockpit solutions for automakers;
• Step 4: Complete selection and configuration of AI-empowered interactive functions on the terminal — Provide a variety of theme libraries, make adaptations to different types of IVIs, and support being equipped on Android / Linux and other systems, so as to cater to more needs of the OEM;
• Upon the completion of all the above-mentioned configurations, a preview version can be released with one touch, and an exclusive SDK will be generated after a 5-minute wait.
As a high-tech enterprise with R&D as its corporate DNA, since its establishment, PATEO has obtained 4,380 intellectual property rights (IPRs), including 2,975 invention patents, accounting for 82%, as well as 849 authorized patents, 445 registered trademarks, and 101 software copyrights. It is worth mentioning that in 2019, it ranked 104th on the list of “China’s Top 500 Companies by Patent”, being the number 1 IoV patent holder among the players in the automobile, auto parts and Internet segments in China. At the same time, according to the “2019 Automotive Patent Innovation Index” released by China Automotive Technology & Research Center (CATARC) at 2019 China Automotive Intellectual Property Annual Conference, it continues to win in a row first place in China and 7th place in the world in terms of intelligent connectivity patent in the IoV & intelligent connectivity segments, together with 26th place among China’s automotive patent holders.
In the future, the company will always be committed to creating a leading AI-powered interaction experience and bringing low-cost but highly adaptive Quick Apps to auto enterprises with both quality and quantity guaranteed, step efforts on seizing more market opportunities for the development of Qing AI voice technologies with technical barriers, assets & platform as strong backing, and look forward to more cooperation with an open & win-win attitude.