Learning Light Transport the Reinforced Way
This addresses efficiency in computer graphics rendering, though it appears incremental as it applies an existing method (reinforcement learning) to a known bottleneck in light transport simulation.
The paper tackled the problem of noisy images in light transport simulation by using reinforcement learning to learn importance sampling, which dramatically reduced the number of zero-contribution paths and resulted in much less noisy images within a fixed time budget.
We show that the equations of reinforcement learning and light transport simulation are related integral equations. Based on this correspondence, a scheme to learn importance while sampling path space is derived. The new approach is demonstrated in a consistent light transport simulation algorithm that uses reinforcement learning to progressively learn where light comes from. As using this information for importance sampling includes information about visibility, too, the number of light transport paths with zero contribution is dramatically reduced, resulting in much less noisy images within a fixed time budget.