From a historically perspective the collapse of many civilizations, including that of Easter Island, has come about as a result of environmental damage. And now the world is approaching a critical juncture of causing irreparable damage to both ecosystems and mankind's communities throughout the global environment. Even while awareness of these problems that threaten our very existence has been growing, under the pressure of solving more immediate issues such as the financial crisis and regional conflicts we have still not been able to unite efforts. What preparations have we, who have been mortgaging the future of the next generation, made for them to survive through the next difficult 20 years?
Clearly, controlling global warming hinges on the question of whether or not we have enough intellect to comprehend methods to achieve that goal. Whether or not we can terraform planets that humans will want to relocate to also hinges on the question of whether we have the intellect to reconstruct the climate of planets like Mars or Venus. Therefore, if we were to create and take advantage of machines with equal or greater intelligence than us - namely artificial intelligence (AI) - we may find some solutions before environmental damage reaches a critical point, by greatly accelerating the progress of science and technology.
However, a long-standing problem that has been a major hurdle blocking the realization of super-human AI is that the machines themselves cannot acquire Knowledge Representation from the real world. Because of this, even global information sharing in the information revolution, which began at the end of the 20th century, has been confined to circulating (mostly linguistic) man-made knowledge. In more recent years, however, deep learning technologies that can learn abstract concepts similarly to the human brain (neocortex) have achieved success by stacking neural network models that imitate the brain's neural circuits to deep hierarchical levels. This represents a revolutionary innovation that will be a breakthrough in facing the above-mentioned challenges. In other words, humankind is now at the brink of a new era aimed at the realization of machines that can gain knowledge from real-world information and autonomously make new creations.
DWANGO believes that a real technological revolution in artificial intelligence is on the horizon, and intends to establish a new Artificial Intelligence Laboratory. DWANGO has set a goal to realize a Japan-pioneered super-human AI that could become a gift to future generations. While using brain-related neuroscientific findings as a point of reference, DWANGO is advancing research with particular focus on a 'Whole-Brain Architecture' research approach as the quickest route to realizing this artificial intelligence. Whole-Brain Architecture seeks to reproduce the computational capabilities of the whole-brain as a combination of machine learning components. To this end, the laboratory will primarily expand and develop deep learning technologies to advance research that aims to acquire a diverse range of knowledge, and to achieve creativity and stability through construction of a computational model of the hippocampal formation, which is where memories reside. The laboratory will also advance research on cognitive architectures for integrating multiple machine learning devices that correspond to parts of the brain concerned with higher orders of intelligence.
While emphasis will be placed on artificial intelligence, coordination with a wide range of academic fields will be needed to realize this research initiative, including the fields of machine learning, neuroscience, cognitive science, software engineering, developmental science, and robotics. Accordingly, DWANGO intends to leverage its unique traits and information sharing capabilities to carry out research in collaboration with a wide range of organizations that will encompass industry, academia, and government, while simultaneously contributing to the creation of a venue in which to carry out these collaborations.
|Title||Dwango Artificial Intelligence Laboratory|
|Established||October, 1st, 2014.|
|Research areas||Artificial Intelligence, Artificial General Intelligence, Whole Brain Architecture|
|Address||Kadokawa Hongo Building., 5-24-5 Hongou, Bunkyo-ku, Tokyo, Japan 113-0033|