Machine learning using Microsoft AZURE and NVIDIA Digits makes machine learning possible from PC to embedded, including server-type data analysis. The knowledge hierarchy is compatible with everything from shallow to deep objects, and the architecture covers a wide range from probabilistic models to neural networks. With this, it is able to fit the service functions necessary to the user.
AI (Artificial Intelligence) complexly combines various machine learning elements to construct an optimum system tailored to the user’s own specific needs.
With a wide range of machine learning used, including Deep Neural Network/SVM/Random Forest/Factorization Machine/boosting and more, We provide optimized services tailored to each respective function (intention estimation/prediction support/Task Agent/independent learning). We also help construct learning data (initial learning/dynamic learning) together with our users.
We provide user comfort through utilization of sensor technology for IoT. First, users' behavior is understood using the automobile, which is a mass of sensors, to provide a comfortable driving space. For cars, user comfort if provided from a very wide range of data that various devices and sensors in the vehicle operate complexly and at high speeds. This is similar to the future environment in which 50 billion terminals will connected in an IoT society, and we carry out testing verification fir automobiles for this society in advance.
For network content, user behavior is analyzed using the car’s navigational system, allowing us to provide all new services.
Also, for network content, we analyze attribute matching among users in order to provide comfortable services. The navigation system learns users’ behavior and provides more even further optimized service. By linking with voice, operability can be improved.