Virus attacks are a fatal threat to all computing systems. Currently, most virus detection methods are based on searching for a virus signature in a suspicious file, to determine if the file contains a specific virus. However, if the signature is mutated, the virus cannot be detected without additional work. This paper focuses on the additional work required. This research is known as clone selection with functional optimization, and is in the area of artificial intelligence (AI).

Since every virus signature must be stored and compared to every computer input file, this process of machine learning is time critical. The actual contribution of this paper is in the improvement of a classical approach with a backtracking known as reverse transcription, with the goal of improving the performance to search for a mature clone.

For beginners in this area, this paper is not a good start, for three main reasons. First, since the authors’ work strongly depends on the work of others, they refer readers to other papers, without discussing them. Second, the convergence of the authors’ method is not discussed. Last, the concept of “searching with backtracking” is only tested with benchmarks for proving function optimizations; the effectiveness of detecting new viruses is not tested. However, that seems to be the status quo on anti-virus systems research.

In summary, this work is a small step toward an ultimate goal of creating an immune system against computer viruses. It is an advanced reference on applied research into anti-virus methods.