In biologically inspired design, designers typically search for natural language documents describing biological systems relevant to their problems. Then they construct an understanding of the biological systems described in the documents for transfer to a given problem. These are difficult, labor intensive and time consuming processes. Thus, we are constructing a virtual librarian called IBID for supporting designers in locating and understanding biology articles relevant to their design problems. IBID first extracts knowledge of the function, the structure, and portions of the causal mechanisms of biological systems from their natural language descriptions. Then, it organizes this knowledge as a Structure-Behavior-Function (SBF) model. Finally, it uses the SBF annotations to retrieve biology articles relevant to design queries. To extract causal mechanisms, IBID uses machine learning techniques to identify portions of a document that describe causal processes.