Technologies Embodied in KIBIT
Background of AI engine KIBITFRONTEO has developed KIBIT, its artificial intelligence (AI) engine, through its services for the uncompromising task of finding evidence in cross-border litigation. This requires us to find data that can be used as evidence from a huge volume of text data within limited time. To meet this challenge, KIBIT learns tacit knowledge obtained by attorneys through experience or intuition and extracts relevant data from a vast amount of text data in accordance with the judgment criteria of attorneys.
However, the data from which KIBIT can learn the judgment criteria of attorneys is not always available in large quantities. What matters then is whether the AI engine can learn from a smaller volume of data. Essentially, texts are ambiguous enough to allow a variety of interpretations, and it is indeed difficult to extract required data from them. In the face of such restraints, FRONTEO has developed a proprietary technology of “landscaping,” or an algorithm to process natural language, which enables the engine to learn from a small volume of data and extract that which is most desirable. This unique algorithm is incorporated in KIBIT.
At present, KIBIT is in operation not only in Japan but also in the U.S., South Korea and Taiwan, among other countries and territories. Furthermore, KIBIT is expanding into other domains of application beyond the initial purpose of finding evidence in cross-border litigation.
How KIBIT got its name
KIBIT is an AI engine independently developed by FRONTEO in Japan. The name KIBIT is derived from the Japanese word “kibi” (meaning subtlety) and “bit,” the smallest unit of information in computing. KIBIT thus incorporates our desire to develop AI that can understand the subtle elements of human thought.
Characteristics of KIBIT
KIBIT is well fitted for text analysis. It features AI-related technologies that perform learning and assessment processes and know-how that has been accumulated and systematized through on-site experiences of data analysis. Built on these strengths, KIBIT can learn, from a small volume of training data, the human subtleties of persons who selected the data or the tacit knowledge grounded in such persons’ experiences and judgment. The underlying technologies are utilized across the full spectrum of FRONTEO’s business domains and incorporated in its product lines, where their capabilities are exercised in a variety of business scenes.