← Back to CoursesArtificial Intelligence: PhD Level

Neuroanatomy Explorer

Drag to rotate · scroll to zoom · click regions to explore

View
Loading 3D model…

Click a region
to explore it

Memory Deck

Flip each card and rate whether you knew it. Your score is saved.

Term
Definition

Deck complete — score saved.

Match the Pairs

Match each term to its definition. Finish the board to earn your score.

All matched — score saved.

Concept Constellation

Every key idea in this course, mapped as an explorable 3D constellation. Drag to rotate, scroll to zoom, click a node.

Click a node to read its definition.

Classical Search, Heuristics, and Adversarial Reasoning

Manual: General · Subject: Artificial Intelligence

Covers uninformed and informed search, heuristic design, and minimax reasoning in adversarial domains.

Classical Search Algorithms

From breadth-first search to A*

Classical search explores a discrete state space to find a goal path. Breadth-first search is complete and optimal for unit costs, while depth-first search uses less memory but can fail to find shallow solutions first. Uniform-cost search generalizes BFS to arbitrary nonnegative costs.

Blind vs informed search

Blind search

  • Uses only problem structure
  • No heuristic guidance
  • Often exponential in practice

Informed search

  • Uses heuristic estimates
  • Can focus exploration
  • May be optimal with admissible heuristics

Heuristic properties

Admissible
Never overestimates the true cost to a goal.
Consistent
Satisfies a triangle inequality across edges.
Dominating
Is greater than or equal to another admissible heuristic everywhere.
💡

Research insight

A better heuristic is not merely more accurate on average; it must also be cheap enough to compute and compatible with optimality guarantees if those matter.

Adversarial search methods

MethodBest use caseKey limitation
MinimaxPerfect-information zero-sum gamesExplodes combinatorially
Alpha-beta pruningGames with good move orderingStill exponential worst case
ExpectimaxStochastic opponent or chance nodesRequires probabilistic modeling
Monte Carlo Tree SearchLarge branching factorsRelies on rollout quality and exploration tuning

Which property ensures that A* tree search is optimal when using graph search with closed lists under standard assumptions?

What is the main advantage of alpha-beta pruning over minimax?

Designing a heuristic for a new domain

  1. 1

    Step 1: Identify relaxed or abstracted versions of the problem.

  2. 2

    Step 2: Derive lower bounds from those simplifications.

  3. 3

    Step 3: Check admissibility and, if needed, consistency.

  4. 4

    Step 4: Benchmark heuristic accuracy against computation time.

  5. 5

    Step 5: Integrate the heuristic into search and measure node expansions.