This section includes my experiences from the Trustworthy Machine Learning course, where I experimented with disrupting the recognition capabilities of deep neural networks using methods like FGSM (Fast Gradient Sign Method) and DeepFool and so on.
Additionally, I worked with two-photon imaging data generated in the lab to model the response patterns of artificial neural networks, which shows the similarity between ANN and brain.
Furthermore, I engaged in training operational models, such as inverted pendulums, using reinforcement learning techniques. These projects not only enhanced my understanding of machine learning's vulnerabilities and robustness but also provided practical insights into the application of neural networks and reinforcement learning in modeling complex systems and behaviors.
In my Data Structures and Algorithms course, I studied dynamic programming, linear lists, stacks, heaps, trees, and graphs. The final project involved creating a robot capable of competing within a specific type of graph. Our bot performed very well.