GenAIvation 2025: Deep Dive into Neural Networks

On Day 4 of the GenAIvation 2025 bootcamp, organized by the Departments of Information Science & Engineering (ISE) and Computer Science & Design (CSD) at Srinivas Institute of Technology, the focus was on advancing participants’ understanding of neural networks. The session explored Single-layered and Multi-layered Perceptrons, introduced the concept of Error Backpropagation (EBP), and detailed the Forward and Backward Propagation processes. Students also learned about key optimization techniques including Momentum and Loss Functions (Cost Functions). Further, the session addressed Multi-class Classification problems and the use of the Softmax Activation Function. Challenges like the Vanishing Gradient Problem were discussed, with solutions such as Dropout and Regularization Techniques explained to enhance model performance. Participants appreciated the concise yet effective teaching approach, reinforced with hands-on coding exercises. The event was supported by Dr. Shrinivasa Mayya D, Principal of SIT, who emphasized the relevance of such initiatives. Dr. Praveen Shenoy, IIC Coordinator, ensured alignment with the IIC’s mission, while Prof. Sudarshan K, Head of ISE, provided constant support. The event’s smooth execution was thanks to Prof. Rakesh Malya and Prof. Sowmya, whose meticulous coordination provided an enriching experience for all participants.