ACM Transactions on Information Systems, 1996
Guglielmo, Eugene 19nn – 20nn
Curiosity. Before Google, Yahoo, Bing, the World Wide Web (WWW), IBM Watson, the US Navy needed to search and find multimedia information – audio, video, graphic, sounds, and images using keywords and natural language statements. There were potentially millions of files that needed to be indexed and searched with a high level of recall and precision in order to prepare the sailors and troops for the next mission within seconds. What technologies could be used to accomplish this?
Invention. This problem was the focus of a doctoral research conducted at the Naval Postgraduate School resulting in a system called Epistemological Information Retrieval Applied to Multimedia (MARIE). The architectural components are shown in the figure above. Coarse (keyword) and fine-grain (semantic) searches guided by a type hierarchy were used to facilitate the indexing and search. Distributed workstations were interconnected to parallelize the search. The results of the search were displayed in a graphical window as shown in the diagram below. Sound familiar? This research predated the WWW and all of the modern search engines by at least five years and was funded by the US Navy. Implementation. Prolog, C, Solaris, Oracle, TCP/IP.
