Global Access to Multilingual Information

This summer I helped to bring together a group to dis cuss something that we called Multilingual Informa tion Discovery and Access We shared a belief that the topic was in some sense larger than the prob lem of cross language information retrieval that has mo tivated the work of many of those who participated and towards that end we sought to include researchers from a broad range of related disciplines That experi ence led me to re ect on how my own perspective has evolved over time and where I think the eld is headed This International Workshop on Information Retrieval for Asian Languages is a particularly apt forum in which to present these thoughts since for the past year our research team at Maryland has been working to under stand whether techniques we have used for European languages can also be applied to Asian languages such as Chinese and Japanese I hope that through this dia log we can build close ties that will prove helpful as we work on problems that are of mutual interest I am not sure that I can distinguish meaningfully between discovery and access so in what follows I will refer to the topic at hand as multilingual informa tion access Aspects of multilingual information ac cess go by a variety of names in a broad range of elds In information retrieval cross language retrieval and information ltering are studied In arti cial intel ligence specialists in what is variously known as com putational linguistics natural language processing or language engineering study machine translation in formation extraction and text summarization while others study ontologies machine learning and tex tual data mining Researchers in library and informa tion science are interested in some similar problems in cluding indexing languages machine assisted index ing automatic abstracting and selective dissemina tion of information In human computer interaction localization and information visualization are key issues and the World Wide Web community has a spe cial interest group working on what they refer to as in ternationalization Modalities other than character coded text are important as well speech processing re searchers are working on topic detection and tracking and in the eld known as document understanding the application of similar techniques to scanned page images is also of interest In digital library research multilin gual metadata is a closely related topic And related so cial issues such as information access international information ow and di usion of innovation are of interest to researchers in a broad range of elds includ ing sociology economics journalism and information science