On Scientific Discovery–A Decision-Machine Approach

The traditional approach to scientific discovery has strictly followed an entity – property – structure –> function –> numerical value one, an approach that has sought to seek out the "terminal" structure of nature; we've developed a decision-machine that sets out conducting scientific discovery "following" an event – information – operation –> matrix –> vector approach with an emphasis on simulating nature's behavior. By defining three induction rules for the decision-machine to obey: 1) similarity degree, 2) effective operational level, and 3) consistency, the decision-machine is able to ensure the effectiveness of inductive reasoning. This way, the decision-machine has a slight advantage over human scientists in three areas: being able to keep a "cool head" and not be influenced by emotions (objectivity), not being dragged into endless debate of whether potential hypothesizes are viable (effectiveness), and being able to work around the clock nonstop continually self-learning and self-adapting (automation). The decision-machine is constructed by drawing upon the concept of quantum superposition and Darwinian natural selection; superposition is used to generate all the possibilities of the real world (mother nature's choices), then natural selection is utilized to evolve the "fittest" hypothesis that most approximately conforms to the real world, and lastly through means of recursive learning, the decision-machine is able to dynamically “deal with” the unknown of the future.