Reliable AI that Develops Society
Deep Learning is one of the most representative technologies in recent AI and shows high performance in pattern recognition and analysis. However, as it cannot explain the reasons for its judgment, it is called "black box AI." Due to this restriction, it is difficult to apply AI to the fields requiring high reliability and persuasiveness such as healthcare, finance, and corporate management that especially need important decision-making.
Fujitsu Labs has developed the world's first machine learning technology called "Deep Tensor" that can directly analyze the relations among numerous pieces of real-world data ranging from intercompany transactions to material structures. We also developed a technology for building a large-scale knowledge base, which is called a "knowledge graph" and consists of vast knowledge existing around the world such as academic papers, by using our unique technology. We combined these two technologies and developed the world's first technology that enables AI to explain the reasons and basis (evidence) for its judgment by constructing a logical path from input to the inference result of AI, which can be used by people securely. With this technology, we can realize explainable AI that overcomes the restriction of ordinary deep learning and can be used by people with high confidence. We are able to provide the information on unknown causal relations and academic papers supporting their relations to genomic medicine specialists, by using knowledge graph consisting of the data stored in the open databases of life information science and the data in more than 10 million medical documents. We are trying to realize individual medicine optimized for each patient and find a new treatment.
In addition to the above activities, we are also promoting advanced research on AI such as a new machine learning algorithm, mathematical approach to real world problems, and natural language understanding technology.