Speaker:Ibrahim Mesecan

Finding and fixing software faults is a challenging and expensive process. Companies spend billions of dollars on this every year. To reduce this cost, many tools have been created to automatically detect and fix software faults. Automated program repair (APR) and Genetic Improvement (GI) are now widely studied and have been used in industry. However, common APR/GI tools make several assumptions. For example, they assume that there is a fault somewhere in the target program and that it was previously detected, giving us a set of both failing and testing test cases. And many common genetic mutation operators assume that there is a solution somewhere in the target program. However, these assumptions may not hold for new domains. For example, information can leak from a program without leading to a functional fault. And for non-traditional programming domains, the program repair paradigm may not have a direct mapping, making it difficult to utilize. In this research, we propose a framework for adapting APR and GI to work in new domains. We demonstrate the feasibility of this framework on two domains where APR and GI have not been previously used. First, we apply automated program repair on molecular programs written in the form of chemical reaction networks. Then we apply genetic improvement on programs with information leakage, a security flaw. Preliminary results find correct patches for 59% of the molecular programming subjects. We are also able to reduce information leakage while retaining functionality of patches 42.6% of the time, however we see the need for defining this as a multi-objective optimization problem.

Committee: Myra Cohen (major professor), James Lathrop, Robyn Lutz, Wensheng Zhang, and Philip Dixon

Join on Zoom: Please click this URL to start or join. https://iastate.zoom.us/j/97655812144 Or, go to https://iastate.zoom.us/join and enter meeting ID: 976 5581 2144