Turkish cultural heritage and digital archives are being transferred to the native AI BİLGE
Digital library and data pool studies are underway to ensure that accurate information about Turkish history and literature is included in native artificial intelligence systems.
A digital library and a qualified data pool are being established to ensure that accurate information regarding Turkish history, literature, and culture is included in artificial intelligence engines. It was reported that digital archives such as the Dictionary of Turkish Literature Names and the Dictionary of Turkish Literature Works will serve as sources for BİLGE, the native artificial intelligence model developed by TÜBİTAK BİLGEM.
Preparations for a digital library, carried out to ensure that artificial intelligence engines contain accurate information regarding Turkish history, literature, and culture, are ongoing. A qualified data pool aimed at enabling researchers to easily access sources and provide data to artificial intelligence systems has been created. It was stated that this accumulation will serve as a source for BİLGE, the native artificial intelligence model developed by TÜBİTAK BİLGEM based on the linguistic structure and cultural accumulation of Turkish.
Speaking at the 15th International Communication with the Future Workshop held in Skopje, Prof. Dr. Mustafa İsen pointed out that a significant portion of the rich biographical accumulation of the Turkish and Islamic worlds has not yet been digitized. While the Dictionary of Turkish Literature Names, prepared under the leadership of Ahmet Yesevi University, includes the biographies of approximately 14,000 poets and authors, the Dictionary of Turkish Literature Works contains 8,085 work entries. These resources provide access to literary information belonging to different centuries from a single digital environment.
For digitized cultural heritage to be utilized by artificial intelligence, texts need to be organized in a format that computers can process. Through this process, it is aimed that artificial intelligence answers questions about historical figures and works from reliable sources and displays the works it is based on. The BİLGE model family, developed by TÜBİTAK BİLGEM, is designed in various sizes ranging from one billion to 122 billion parameters, aiming to correctly understand Turkish and interpret cultural expressions.