Gaussian Process Based Optimization

The project demonstrates optimization of hyper-parameters using Bayesian Global Optimization.

Details:

• Convolutional Neural Network (3Conv2D + 3Dense)
• Street View House Number Dataset: Link
• Hyperparameters under consideration .. 1.Starting Learning Rate .. 2. Learning Rate Decay .. 3. Mini-batch size .. 4. Dropout parameter p1 for first fully connected layer .. 5. Dropout parameter p2 for second fully connected layer

Tasklist:

• [X] Boilercode
• [X] Mid-term Report
• [ ] Implement Bayesian Global Optimization: Link
• [ ] Hyperparameter tunig
• [ ] Performance Analysis

References:

1. J. Snoek, H. Larochelle, and R. P. Adams, “Practical bayesian optimization of machine learning algorithms,” in Advances in neural information processing systems, 2012, pp. 2951–2959
2. K. Weinberger, “Bayesian global optimization,” http://www.cs.cornell.edu/courses/cs4780/2018fa/lectures/lecturenote15.html,2018.

Contributors:

• Ankush Tale
• Shubham Shah
• Amey Athale
• Uday PB