Direct Optimization CSC2547 Adamo Young, Dami Choi, Sepehr Abbasi - PowerPoint PPT Presentation
Direct Optimization CSC2547 Adamo Young, Dami Choi, Sepehr Abbasi Zadeh Direct Optimization A way to obtain gradient estimates that directly optimizes a non-differentiable objective. It has first appeared in structured prediction
Gumbel Process
Gumbel Process We know:
Gumbel Process We know: Therefore:
Gumbel Process We know:
Gumbel Process A B
Gumbel Process A B
Gumbel Process A B
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value.
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value.
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value.
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3 1.3
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3 1.3
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3 1.3 1.1
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3 1.3 1.1
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3 1.3 1.1
Trajectory Generation ● Lazily create partitions of trajectories. ● Recursion rule: ○ For , copy parent node’s value. ○ For the remaining choices of actions, group them and compute truncated value. 1.3 1.3 1.1 1.3
Recommend
More recommend
Explore More Topics
Stay informed with curated content and fresh updates.
Sambuz