Our research group (NLP&CL) at KAIST Computer Science department is interested in utilizing natural language processing techniques for biomedical applications, where our activities are currently focused on the following areas: information extraction for biomedicine, automatic construction of gene and protein ontologies, visualization of inferred gene and protein networks, natural language query interpretation for heterogenous database access, automatic creation of gene summary, and prediction of RNA secondary structure with grammars. These applications can be synergistically integrated for knowledge discovery. For example, various pieces of information extracted from biomedical literature are useful to extend ontologies and, in return, the extracted information can be classified by looking upon the extended ontology. The extracted information can also be a source information of visualization, and, in return, the visualization gives feedback from users back to the information extraction systems. An ontology provides a structural knowledge base with which visualization systems display protein pathways in a well-structured form, and, in return, visualization gives feedback on the knowledge base from users.