Informace o projektu

Dynamic Learned Indexing for Similarity Searching (Dynamic Learned Indexing for Similarity Searching)

Kód projektu
MUNI/IGA/1071/2021
Období řešení
1/2022 - 12/2022
Investor / Programový rámec / typ projektu
Masarykova univerzita
Fakulta / Pracoviště MU
Fakulta informatiky

The project's aim is to create an efficient dynamic high-dimensional index for processing complex and unstructured data that are subject to frequent changes (INSERT, DELETE operations) with the use of machine and deep learning methods. It builds on the latest research which favors the use of learned approaches as opposed to rigid data structures for index construction. The proposed index's performance will be verified on synthetic as well as real-world image and motion capture datasets.

Používáte starou verzi internetového prohlížeče. Doporučujeme aktualizovat Váš prohlížeč na nejnovější verzi.

Další info