First International Workshop on the role of Semantic Web in Literature-Based Discovery

Printer-friendly version
Thursday, October 4, 2012 - 12:00am to Sunday, October 7, 2012 - 12:00am

Literature-Based Discovery (LBD) is characterized by uncovering hidden but novel information implicit in non-interacting literatures. The notion of LBD was first proposed by Don R. Swanson over two decades ago when he postulated that two concepts (A,C) may be logically related through some intermediate concept (B), common to seemingly disjoint literatures. This seminal idea has largely influenced efforts towards LBD automation in the biomedical domain and LBD continues to be an integral part of the evolution of biomedical science. Primarily, LBD has been instrumental in supplementing and guiding scientific experiments that lead to innovations in diagnosis, treatment and preventions mechanisms.

Much of the early LBD research however, relied almost entirely on Information Retrieval (IR) techniques, such as term and concept co-occurrence, to uncover unknown associations in the large volume of scientific literature now publicly available. Only recently has significant attention been devoted to semantics-based techniques that leverage Semantic Web technologies to exploit the meaning of associations between concepts to facilitate LBD. While generally more intuitive than IR techniques, the feasibility of semantics-based approaches has not been fully demonstrated. Many challenges still exist. Some of these include:

  1. Fine-grained extraction of semantic information (called semantic predications) from text corpora.
  2. Extraction and identification of meaningful (semantic) associations between concepts. Such associations are typically represented by paths in large data graphs.
  3. Achieving scalability given the combinatorial explosion that arises when traversing large graphs. The overwhelming number of edges between concepts increases the complexity of graph traversal and compounds the difficulty of finding relevant associations.
  4. Developing techniques for clustering, aggregating, and analyzing extracted semantic associations for sense making, question answering and ultimately LBD.
  5. The semantic integration of information expressed in text corpora with background knowledge.
  6. By engaging researchers from both the Semantic Web and LBD communities, we anticipate an exchange that will facilitate the advancement of LBD by exploiting available Semantic Web resources. Researchers are encouraged to submit original manuscripts on the application of Semantic Web technologies, representations and techniques to Literature-Based Discovery.