M.Sc. C.S. Thesis Defense: Prometheus Peter L. Lazo (Spiking Neural P Systems with Stochastic Application of Rules)

Zoom Link https://up-edu.zoom.us/j/87005325265

Meeting ID: 870 0532 5265

Meeting Password: 12329827

This work continues the investigations of introducing probabilities to spiking neural P systems, SN P systems in short – membrane computing models inspired from biological spiking neurons. A particular interest for SN P systems in this work is the nondeterministic selection of applicable firing rules. Rules represent the possible reactions of a neuron to the number of electrical impulses, or spikes, present. Intuitively, having nondeterministic selection can be interpreted as having a random choice with equal probabilities for all options. This seems unnatural in some biological sense since some reactions are more active than others in general as emphasized in Obtulowicz, A., & Păun, G. (2003). (In search of ) probabilistic P systems. BioSystems, 70(2), 107-121. This work proposes SN P systems with stochastic application of rules, *SN P systems in short. *SN P systems are variants which introduce a stochastic process a priori to the application of rules in SN P systems. Results found that the stochastic process introduced to the nondeterministic selection of firing rules covers rule application in general. In effect, application of single firing or forgetting rules also became stochastic. *SN P systems are proven to be computationally universal. However, when simulating a register machine, rule application probabilities of 1 were used in some neurons. When setting rule application probabilities strictly to < 1, the neurons became asynchronous. The use of extended rules are introduced to *SN P systems to address the asynchronous behavior. This study then also proves that the asynchronous *SN P systems using extended rules are computationally complete. Lastly, this study compares and contrast *SN P systems with the classic SN P system, stochastic SN P variants, and SN P variants that use extended rules.