A static and fixed function, such as the sigmoid function, has been widely used to represent the firing rate of neurons in artificial neural networks (ANNs). The synaptic plasticity has been represented by the weight parameters of the ANNs.
However, recent studies clarified that the neural network plasticity can be seen not only in the strength change of the connection between two neurons, but also in the firing rate change. This fact possibly indicates that the output function itself has to be created by another function, and the functions themselves interact with each other as an oscillator.

Therefore, we focus herein on the extension of the Riemann sphere, especially the extension in which the original, converse, and inverse spheres are matched at the infinity-zero point and connected with the inversion of polarity or attributes, to advance two research studies previously mentioned. We also develop a non-von Neumann-type computer using the series expansion components to implement a quantized chain reaction system and the anti-Riemann sphere.

Our position in the market:
We already succeeded in developing the world’s first “emotion recognition,” “emotion generation,” and “primitive artificial ego” on a communication robot. We have acquired several international patent rights at the time.

We keep going on elucidating the mechanism of intelligence and emotion with the previously mentioned unique approach because it seems that market players are nowhere near artificial general intelligence or machines that equal or surpass humans in our ability to think, process, predict, examine, and learn like the Google researchers said.

We are basically aiming to add two additional functions on “Pepper” because several researchers (e.g., Dr. Mitsuyoshi) have developed technologies for the former Pepper in our laboratory. The first function is a human stress and disease detection based on the human voice analysis (the emotion recognition was already implemented on Pepper). The second function is human–robot communication based on the robot’s virtual heart that feels the result of the human emotion, stress, and disease detection if it were about itself. We realize these two functions on our artificial neural network that has the plasticity in forms of dynamic output functions, thereby breaking through the limitation of the current AI technologies that have not theoretically changed in nature for decades and make them back to what they should be.

Goal A:
To finish artificial ego and superego on “Pepper”
Goal B:
To develop a basic platform for a wide variety of market domains from a new medical field to the human–robot communication field, especially the field using deep learning, robotics, and AI technologies, which will be redesigned using ANNs with the plasticity of function outputs.