Philosophers argue that scientific discovery is far from being a rule-following procedure with a general logic: More likely it incorporates creativity and autonomy of the scientist, and probably luck. Others think that discovery can be automatized by some computational process. Based on a concrete example of Schmidt and Lipson Schmidt and Lipson (2009), I argue that the bottom-up discovery is computable and that both aspects of creativity and autonomy can be incorporated. The bio-inspired evolutionary computation (genetic algorithms) are the most promising tool in this respect. The paper tackles the epistemology of applying a evolutionary computational and genetic algorithms, to the process of discovering laws of nature, invariants or symmetries from collections of data. Here i focus on more general aspects of the epistemology of evolutionary computation when applied to knowledge discovery. These two topics: computational techniques applied in science and scientific discovery taken separately are both controversial enough to raise suspicions in philosophy of science. The majority of philosophers of science would look with a jaundiced eye to both and ask whether there is anything new to say about discovery and computers in science. This paper is a first stab to the philosophical richness of computational techniques applied to the context of discovery. I discuss the prospect of using this type of computation to discover laws of nature, invariants or symmetries and appraise their role in future scientific discoveries.