We are exploiting four types of AI technologies that make up corevo®.
Our aim is to support the creation of a safe, secure society in which people can find fulfilment.
Agent-AI understands a person's intention or feeling from his/her voice, speech or body language and is able to engage in an advanced [conversation through a robot or a virtual agent that performs multimodal interactions, including gestures.
It enriches a person's everyday life by "living" alongside with the individual, supporting or actually taking over certain activities, with each party performing the activities it is better at. Looking forward, we aim to develop AI that is the lifelong companion of the user, thereby realizing a society in which both young and old can lead fulfilling lives.
· Agent-AI understands a person's intention and surroundings using microphones, cameras and sensors.
· It sorts out (structures) the real world by analyzing and understanding the wide spectrum of information around us, and creating knowledge from these activities.
· It provides highly expressive and realistic interactions using multimodal communication and multiple devices.
· Agent-AI supports or substitutes for routine tasks, such as those in contact centers and over-the-counter services in financial institutions.
· It supports the tasks performed by medical doctors and care workers, thereby relieving them of some of their workload.
· It supports and enhances the ordinary lives of the elderly or parents raising small children who experience various forms of inconvenience in terms of everyday mobility or communication.
making it possible to support people in improving their sports skills or mental wellbeing.
Heart-Touching-AI understands, acts on and expands a person's essential, fundamental and yet subconscious elements, instinct and physical body.
It removes mental and physical barriers between individuals or between individuals and their environments or society, thereby realizing a human society that is not just "convenient" but "comfortable." Furthermore, in sports and the arts, Heart-Touching-AI enables anyone to have the kinds of special experiences that only professionals are said to have.
· Heart-Touching-AI is realized by combining advances in cognitive neuroscience/language learning science with information processing technologies, such as searching and language processing.
· It uses non-invasive bioelectric sensing technology to determine a person's mental state, the activity of the autonomic nervous system and muscles in real time, and acts in response to the cognition and movements of the individual.
· It compensates for or expands a person's mental and physical functions through operations of robots or computers that work on information that is not consciously perceived by the individual.
· Heart-Touching-AI provides information, supports activities and provides a sense of belonging in a place in a way adapted to a person's mental and physical conditions at the time.
· It acts by responding to the reflexive actions of the body, enabling the user to improve sports skills at the subconscious level. It can transmit a technique that cannot be conveyed with words.
· It provides education support and rehabilitation through a process adapted to the characteristics and growth of each individual.
· It helps to make possible a medical environment, social systems and mental fitness services that heal both mind and body.
Ambient-AI interprets information from sensors in real time. It is geared to the coming ICT society in which a wide variety of devices equipped with sensors and/or actuators are connected to networks. It performs spatiotemporal analysis and deduces cause-and-effect relationships from this analysis. This makes it possible to detect predictive signs well in advance, to explore optimal scenarios, and to design and perform smart guidance and control.
It uses advanced and high-performance data analysis technology based on machine learning in order to solve social problems and create innovative IoT services.
· A real-time cyber-physical system that repeats a cycle of analysis, exploration and design.
· Prediction of near future through spatiotemporal multidimensional analysis
· Deduction of statistical cause-and-effect relationships that can explain the observed phenomena.
· Technology for detecting predictive signs of rare events.
· Proactive control based on learning-type multi-agent simulation technology.
· Stable control of a variety of social/industrial infrastructures based on detection of predictive signs and deduction of cause-and-effect relationships, and reduction in environmental loads · Real-time smart-city management based on spatiotemporal prediction of human and vehicle traffic flows and proactive guidance.
Each Network-AI, which is connected to a network, can connect itself to one of a number of resources and work with it. The network itself works as a single AI, performing both globally optimized control and area-dependent control. Network-AI provides value to customers irrespective of the protocol and the network operator used.
In addition, the application of AI to a network makes it possible to provide a safe, secure and high-quality network to customers irrespective of whether the user is in normal or emergency situations.
· As a number of AI technologies get interconnected in an organic way, Network-AI continues to grow and globally optimizes social systems.
· Exploiting a huge amount of environmental information (network information, weather, social conditions, etc.) sent from network devices, Network-AI provides intelligence (autonomy) to automatic protection, maintenance, design and control of a network.
· With a view to utilizing volatile Big Data safely, Network-AI carries out intra-area statistical processing to prevent the dissemination of privacy information.
· Emergency or quick-response control of a drone, etc. appropriate to the conditions of the area affected by an accident.
· Re-design and re-optimization of the delivery or production plan of a social system by making different AI capabilities share needed knowledge extracted from accident information about the affected site.
· Automatic derivation of the guidelines for control and expansion of wireless and core networks to reflect the forecast of network demand.
· Stable protection and maintenance of a network using verifiable machine learning.