Model building and extension is a labor-intensive process. Machine reading tools can aid this process by assembling information from the vast amount of research literature available. However, the amount of information available is often greater than the modeler's need. VIOLIN (Validating Interactions Of Likely Importance to the Network) was built to address this problem, automatically and quickly comparing machine reading output to a baseline model, and returning judgements to help the user select the most useful Literature Extracted Events (LEEs) for their needs.
Documentation
VIOLIN can be downloaded from TBD
Documentation for VIOLIN can be found at TBD
Description
VIOLIN takes machine reading output and a baseline model, and compares each LEE to a baseline mode. The LEE is compared and judged based on four metrics, and judgements are represented by numerical values which are then calculated into a final score. The output can be used with respect to this total score only, or with respect to it's main classification category: corroboration, extension, contradiction, or flagged. Details on these classifications can be found in the documentation and publications.
Publications and Citations
Publications
- C. E. Hansen, J. Kisslinger, N. Krishna, E. Holtzapple, Y. Ahmed, and N. Miskov-Zivanov, “Classifying Literature Extracted Events for Automated Model Extension,” arXiv preprint arXiv:2110.10841, Accepted at the 13th International Workshop on Bio-Design Automation (IWBDA21).
Citations