Deep learning architectures have revolutionized numerous fields, achieving state-of-the-art results in tasks ranging from image recognition to natural language generation. Amidst this progress, researchers constantly explore novel approaches to enhance the potential of deep learning. In this context, AROM168 emerges as a promising deep learning sy