2D Density Estimation with 10,000 Nodes DDN

This GIF demonstrates the optimization process of 2D probability density estimationn using a Discrete Distribution Network with 10,000 nodes.

• Left image: All samples that can currently be generated
• Right image: Target probability density map
• For demonstration purposes, the target probability density maps switch periodically. Names and sequence of target probability maps:
  • gaussian -> blur_circles -> QR_code -> spiral -> words -> gaussian -> uniform -> gaussian (same at beginning and end, completing a cycle)
• Therefore DDN continuously optimizes parameters to fit new distributions
• Optimizer: Gradient Descent with Split-and-Prune
• The experiment code is in sddn/toy_exp.py, and the experimental environment is provided by the distribution_playground library, feel free to play with it yourself
• Tips: You can use a phone camera to scan out the QR code distribution fitted by DDN at a suitable screen distance