The past two weeks in ISM have been both productive and exciting as I worked on improving my AI model and preparing for the upcoming showcase. My focus has been on refining my original work by incorporating a new technique, which is building on a pre-trained image model with my dataset. This improvement increased my accuracy, allowing me to exceed 70%, which felt like a significant milestone. I am continuously exploring new strategies to push that accuracy even higher, such as data augmentation and fine-tuning parameters. I have also devoted considerable time to preparing my ISM display board for the showcase. Despite the abstract nature of my AI model, I have included visuals to illustrate key aspects of its development, such as accuracy trends and statistics from the training process, as well as a sample dataset image to make the project more accessible to viewers. I have also been carefully crafting the content of my board to highlight the real-world impact of AI in early skin cancer detection and to provide an overview of the technologies I used, such as TensorFlow and various machine learning algorithms. Additionally, I plan to present my laptop with annotated code and a simplified walk-through so that even those unfamiliar with the technical aspects can grasp the steps involved in developing my project. I am excited to showcase my progress and hope that this setup will effectively communicate my research, AI model, and its potential impact to attendees while making the experience informative and engaging!
