A flexible architecture for collaborative information retrieval
Jian‐Yun Nie, Jean Vaucher, Nicolas Anquetil
Abstract
Jian‐Yun Nie, Jean Vaucher, Nicolas Anquetil
Abstract
Many Information Retrieval techniques have been developed, but none is perfect in every case. Moreover, these techniques often differ markedly in their operation principles making them difficult to integrate in a collaborative way. The MARS system (Multi-Agent Retrieval System), was built to test the integration of multiple retrieval methods in a single system in order to offer users more flexible and complementary retrieval tools. The retrieval methods integrated in MARS include not only the classical methods, but also more intelligent ones such as retrieval based on natural language processing and conceptual browsing in a concept space generated automatically from analysis of the corpus. Each retrieval method is implemented via a retrieval agent which may operate either independently or in collaboration with others. This system concerns local retrieval as opposed to global retrieval offered by WWW, Mosaic and Wais; but it may be directly integrated in the latter to enhance access to pertinent data.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Many Information Retrieval techniques have been developed, but none is perfect in every case. Moreover, these techniques often differ markedly in their operation principles making them difficult to integrate in a collaborative way. The MARS system (Multi-Agent Retrieval System), was built to test the integration of multiple retrieval methods in a single system in order to offer users more flexible and complementary retrieval tools. The retrieval methods integrated in MARS include not only the classical methods, but also more intelligent ones such as retrieval based on natural language processing and conceptual browsing in a concept space generated automatically from analysis of the corpus. Each retrieval method is implemented via a retrieval agent which may operate either independently or in collaboration with others. This system concerns local retrieval as opposed to global retrieval offered by WWW, Mosaic and Wais; but it may be directly integrated in the latter to enhance access to pertinent data.
Key concepts: Computer science, Information retrieval, Human–computer information retrieval, Vector space model, Data retrieval, Cognitive models of information retrieval, Document retrieval, Adversarial information retrieval