Speech search

Gareth J.F. Jones
2014 Proceedings of the 37th international ACM SIGIR conference on Research & development in information retrieval - SIGIR '14  
Speech search is concerned with the retrieval of spoken content from archives of speech or multimedia materials. Importantly speech search does not reduce to application of text information retrieval to speech recognition transcripts. Although text information retrieval techniques form a core technology for speech search, speech search generally involves confronting issues less likely to arise in the text domain, such as high levels of noise in the data and the absence of clearly defined units
more » ... f retrieval. This tutorial will provide researchers in information retrieval with an introduction to the challenges and technologies of spoken content search. While speech search has been an active area of research for around twenty years, during much of this time activities have focused on retrieval from collections of well defined and clearly structured content such as radio and television news. The tutorial will review this establsihed work based on the detailed descriptions contained in [4] . The current rapid growth in recordings of more informal audio content, e.g. in educational and enterprise settings. is creating new opportunities and demands for speech search technologies. Despite the developments in existing speech search technologies, these types of data pose challenges for which the solutions proposed to date are either incomplete or require extension or new methodologies. The tutorial will draw on current research to demonstrate the retrieval challenges posed by this content, and introduce work which aims to develop effective solutions, e.g. for search of internet video archives [3], meetings [2] and lectures and presentations [1] .
doi:10.1145/2600428.2602293 dblp:conf/sigir/Jones14 fatcat:wegvdptefnftnmbgnc3dx5bxuu