KAIST develops vector database that preserves AI search accuracy as data changes

Researchers at KAIST have developed a vector database technology that preserves the search paths leading to needed information as records are added and deleted. This allows AI systems to find up-to-date information more accurately and quickly even when older material has been removed.

• Technology preserves search paths to data as records change
• Addresses loss of needed information after deletions
• Improves accuracy and speed of retrieving current data