Multimodal Interaction System for the communication with people
Quantum Artificial Intelligence
The future potential of quantum computing has been highly anticipated in recent years. Following that path, we are focusing our efforts on quantum artificial intelligence. We are convinced that this scientific challenge has the potential to impact traditional human-like decision making models.
We believe that this engagement will soon become a reality, and therefore plan to upgrade our A.I. Galleria engine accordingly. For example, there are research reports that prove that organisms use the quantum effect in photosynthesis, high-precision gene copying, and long-distance moving by using geomagnetism. We think that there is no reason to deny the possibility of developing quantum devices for use in room temperature.
Therefore, we are making our best in developing quantum artificial intelligence to improve performance.
A.I. Galleria is a Multimodal Interaction System meant to facilitate communication between humans and computers.
A.I. Galleria is equipped with an image classification API. It uses Machine Learning to determine the features of an image, and then organises the classified data in the storage of A.I. Galleria.
We developed a robot named AI-Samurai (http://ai-samurai.com) to showcase some of this functionalities. The AI-Samurai uses a learning model helps him recognise people's faces and associate them with their names. Based on this achievement, we are going to develop a mathematical model that can further enhance the communication experience by using additional information available in the images.
A.I. Galleria is equipped with an interaction API. As of May 2016, the supported languages are English and Japanese. This API analyses the contents of the message being transmitted by the user and extracts the intention, then selects a speech strategy and generates a suitable response.
A.I. Galleria can also can analyse the nature of expression being said by each end of the communication channel. By doing this, as the number of interactions increase, the data being generated in the conversation serves as source for refinement and optimisation of the responses. This ultimately allows the system to produce more accurate replies that aim at resembling the conversation held with a person through the Multimodal System
Therefore, we will keep on analysing the data obtained from AI-Samurai (http://ai-samurai.com) World Tour and other demonstration experiments to build an optimal communication interaction model.
In order to generate output, A.I. Galleria uses various mathematical models of image recognition and interaction. The personally specialised mathematical model can be built through the analysis of each model per a counterpart of the communication.
There are several of technologies used, such as deep learning, NLP, graph theory, information geometric approaches, and so on.
In the future, we will continue trying to recreate intelligence within the system of A.I. Galleria, making the development of a quantum artificial intelligence approach, our goal for the quantum theory.