2023, Vol.28, No.6

Journal of The Korea Society of Computer and Information 2023 KCI Impact Factor : 0.65

pISSN : 1598-849X / eISSN : 2383-9945
https://journal.kci.go.kr/jksci

In this paper, we propose a laboratory animal breeding device IVC(Individually Ventilated Cage) system that applies a customized air conditioning control method according to the air information inside the laboratory animal cage. This method maintains an optimal breeding environment in which the experimental animals are not stressed inside the cage by precisely controlling the ACH(Air Change per Hour). In addition, for the convenience of animal experiment researchers, it is a system capable of monitoring breeding environmental variables such as temperature, humidity, differential pressure, ACH, blower RPM(Rotations Per Minute), HEPA(High Efficiency Particulate Air) filter replacement cycle, pre-filter replacement cycle, CO2 concentration, NH3 concentration, and air speed. In order to test the performance of the developed system, two laboratory mice were placed in one cage, and environmental variables were monitored while raising them in the developed system for about 6 hours. As a result of the experiment, the experimental animal breeding environment target value presented in the IACUC(Institutional Animal Care and Use Committee) guidelines was satisfied.

In this paper, we propose an approach to analyze the dynamic LSTM network activation in an LSTM English learner language model trained on a large corpus of English-speaking learners. The objective is to examine the relationship between network activation and performance on the Subject-Verb Number Agreement task. By employing a NA-task probing classifier and conducting ablation experiments, activation patterns are evaluated at each time step. The results reveal a strong link between network activation and the classifier's performance. The proposed model achieved 99.57% accuracy on the evaluation dataset for the NA-task, demonstrating its acquisition of correct grammar rules and accurate prediction ability. To analyze the influence of internal neurons on NA-task processing, specific LSTM neurons are removed and the model's performance is examined. Removing neuron 776 resulted in a more than 10% decrease in performance for plural subjects, while removing neuron 988 led to a 10% decrease in performance for singular subjects compared to the model before removal.

In this study, we propose a methodology for deriving natural additives that expedite cultured meat production through a combination of bioinformatics and artificial intelligence-based approaches in drug development. Leveraging a bioinformatics approach, we identify differentially expressed genes that are genetically associated with muscle and adipose tissue growth in breast cancer cells. Subsequently, employing an artificial intelligence-based drug development approach, we construct a model for predicting protein and compound interactions, and conduct virtual screenings of two natural product databases (COCONUT and the Korean Traditional Medicine Database). This enables us to identify 22 candidates of natural compounds that bind to proteins responsible for inhibiting muscle and adipose tissue growth.