In a recent development, Mr. Sangawa, a M2 student, has released a deep neural network (DNN)-based malware detection model named DeepMalDetect on GitHub. This model is capable of learning complex features through its multi-layered architecture. The repository can be accessed here. This work is based on the research presented at the 2025 International Conference on Intelligent Information Technology (ICIIT 2025), held in February 2025 in Hanoi, Vietnam. The presentation was titled “Parallel-processed Hyperparameter Tuning for Higher Accuracy of Malware Detection.”
By using the publicly available model and source code, experiments utilizing a high-accuracy DNN-based malware detection model can be reproduced. The detection performance was evaluated in terms of accuracy and recall, and the proposed model demonstrated superior performance compared with models presented in prior studies, while using the same number of training epochs.
For further details on the model and experimental methodology, please refer to the following paper.
