March 2020
MacFarlane, Andrew, Sondess Missaoui, and Sylwia Frankowska-Takhari. "On Machine Learning and Knowledge Organization in Multimedia Information Retrieval" Knowledge Organization 47(1)(2020): 45-55. (https://www.ergon-verlag.de/isko_ko/downloads/ko_47_2020_1_e.pdf). - Media and multimedia objects have so many individual features that could potentially constitute some form of metadata that indexing them through artificial intelligence is, in some cases, not possible. Some features are subjective and can only be recognized and assigned by human indexers, but not all. The authors propose a system of “knowledge organization” that begins by sorting identifiable qualities of art, music and video into high-level, middle-level and low-level features. Low-level features "can be extracted using machine learning technologies, whilst high-level features … require the use of human intervention..." (p. 50) Advances in technology increase the possibility of indexing with artificial intelligence, such as graphical processing units (GPU), which "are particularly useful for image processing..." (p. 46) The authors seek to help information professionals see artificial intelligence as "an opportunity rather than a threat," and a "technology to improve the multimedia services they manage." (p. 53) - NN