Tree-of-thought for complex decisions in Agentic AI - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to start a tree-of-thought search with an initial state.

Agentic AI

tree = TreeOfThought(initial_state=[1])

Drag options to blanks, or click blank then click option'

Astart_state

Binitial_state

Croot_state

Dstate0

Attempts:

3 left

💡 Hint

Common Mistakes

Using variable names that are not recognized by the TreeOfThought constructor.

Confusing the initial state with other variable names.

The TreeOfThought class requires the initial_state parameter to start the search tree.

2fill in blank

medium

Complete the code to expand the current node by generating possible next thoughts.

Agentic AI

next_nodes = current_node.[1]()

Drag options to blanks, or click blank then click option'

Aexpand

Bbranch

Cgrow

Dgenerate

Attempts:

3 left

💡 Hint

Common Mistakes

Using method names that do not exist on the node object.

Confusing expansion with evaluation or selection.

The method expand() is used to generate the next possible thoughts or nodes from the current node.

3fill in blank

hard

Fix the error in the code to correctly evaluate the utility of a node.

Agentic AI

score = node.evaluate_utility([1])

Drag options to blanks, or click blank then click option'

Ascore_function

Butility_score

Cevaluate

Dheuristic_function

Attempts:

3 left

💡 Hint

Common Mistakes

Passing a numeric score instead of a function.

Using incorrect parameter names that cause runtime errors.

The evaluate_utility method requires a heuristic_function to estimate the node's value.

4fill in blank

hard

Fill both blanks to select the best child node based on utility and update the current node.

Agentic AI

best_child = max(current_node.children, key=lambda n: n.[1])
current_node = current_node.[2](best_child)

Drag options to blanks, or click blank then click option'

Autility_score

Bselect_child

Cevaluate_utility

Dmove_to

Attempts:

3 left

💡 Hint

Common Mistakes

Using a method instead of an attribute for the key function.

Trying to assign the current node without using the proper method.

The utility_score attribute is used to find the best child, and move_to() updates the current node to that child.

5fill in blank

hard

Fill all three blanks to implement a recursive tree-of-thought search with depth limit.

Agentic AI

def tree_search(node, depth):
    if depth == 0 or node.is_terminal():
        return node.[1]()
    best_score = float('-inf')
    best_action = None
    for child in node.[2]():
        score = tree_search(child, depth - 1)
        if score > best_score:
            best_score = score
            best_action = child
    return node.[3](best_action)

Drag options to blanks, or click blank then click option'

Aevaluate_utility

Bexpand

Cselect_child

Dgenerate

Attempts:

3 left

💡 Hint

Common Mistakes

Mixing up method names for generating children and selecting best child.

Returning the wrong value at the base case.

The function evaluate_utility() returns the node's score, expand() generates children, and select_child() chooses the best child node.