Course overview
So Far
- Agent Systems
- RoboCup
- Other robot architectures
- Three Layer Architectures (planner/Symbolic reactive/non-symbolic reactive)
- Subsumption
- Cognitive Architectures
- Search
- Graph search
- BFS/DFS/BFS/A*/Approx A*/Iterative Deepening
- Logic spaces
- State space planning
- Plan space/Partial Order planning
- Things for search:
- Representation
- Model of World
- Heuristic
- Current State
Yet to see
- Some representations for discrete search
- Building Models
- Building Heuristics
- Relationship to hierarchy
- Incremental/cached/dynamic programming search
- Stochastic worlds
- Search through continuous spaces
- Incomplete state/partial observability
- More on Hierarchy
- Adversarial Agents
Today
- Building Models
- Try actions in different locations
- Learn to predict the results of actions
- Machine Learning problem
- Need to sample entire space to get model over entire space