Planning and reasoning
Compare search, world models, and repeated internal computation.
What you’ll explore
- Algorithms
- World models
- Recurrence
- Search
Your lessons
0 / 21 completeReason with procedures and search
Compare a local shortcut with a procedure that explores alternatives and checks its result.
Use an informed search
Compare uniform-cost search and A* on a weighted map, audit estimates and reopen a location when a cheaper route arrives.
Solve a constraint problem
Build a small timetable, remove unsupported choices and compare backtracking with constraint propagation.
Update beliefs from evidence
Use sensor frequencies to distinguish detection from posterior probability and test how base rates and repeated evidence change a belief.
Factor a probabilistic model
Build an eight-row joint distribution from local tables, test conditional independence, and separate observing a variable from intervening on it.
Infer state across time
Track a hidden moving object by alternating motion predictions and noisy sensor updates, while keeping a distribution over possible positions.
Approximate a belief with particles
Propagate, weight and resample possible positions. Compare a seeded particle filter with exact inference and diagnose lost diversity.
Reason about an opponent
Play a small stochastic game, reveal its search tree and compare adversarial planning, probabilistic planning and a policy learned from sampled returns.
Ask what would change under intervention
Change a causal graph, compare observed and randomized outcomes, and discover why a population effect does not determine an individual counterfactual.
Predict what an action will change
Learn transitions from observations, compare imagined and actual routes, and test when replanning helps.
Refine an idea over several passes
Watch distance information spread through a grid and discover why stable internal states can still be wrong.
Test the parts of a planning system
Compare direct action, search, and refinement under a shared planning budget, then test unfamiliar layouts.
Learn continuous dynamics
Collect cart transitions, fit a next-state predictor, and compare it with simpler baselines.
Know what a prediction leaves uncertain
Separate noisy observations, multiple physical outcomes and disagreement across fitted models.
Small errors become different futures
Compare fresh-state predictions with a model that must live with its own earlier mistakes.
Plan briefly then observe again
Use a fitted model to choose actions, execute one, then replan from the observed result.
Predict in a compact representation
Learn a small visual code and test whether predicting that code retains useful action information.
Remember what the current view hides
Train recurrent state and compare information available during an occlusion.
Improve a policy using imagined experience
Fit dynamics, optimize actions inside that model, then test in the external simulator.
Compare families of world models
Separate generated observations, latent predictions, action conditioning and policy learning.
Test whether the model helps action
Freeze a model and planner before comparing prediction, control and changed-dynamics outcomes.