Min Woo Kim

Research Experience

• Collaborated with EECSL (a lab at Korea University) and other institutions to acquire data for model training .
• Used Python and PyTorch to develop an Attention-Based Fusing CNN-LSTM Hybrid Model , achieving 99.5% accuracy and 97% F1-score in diagnosing 7 motor fault classes using two sensor modalities .
• Interpreted and analyzed research data to draw conclusions on outcomes to further develop the research.
• Reimplemented state-of-the-art (SOTA) models for performance comparison in controlled settings, validating improvements over existing approaches.
• Currently writing a research paper based on this work.
• Applied critical thinking skills to identify potential areas for further investigation within the research topic.
• Engaged in research on Large Language Model (LLM) compression , extensively using C/C++ and leveraging SOTA research and tools .
• Supported fellow researchers with technical assistance, fostering a collaborative work environment.
• Presented research findings at university conferences, effectively communicating complex concepts to diverse audiences.