Why complex tasks need planning in Agentic AI - Challenge Your Understanding

Challenge - 5 Problems

🧠 Conceptual

intermediate

Why do complex AI tasks require planning?

Imagine you want a robot to clean your entire house. Why is it important for the robot to plan its steps before starting?

A Because planning lets the robot skip cleaning some rooms randomly.

B Because planning helps the robot decide the best order to clean rooms efficiently.

C Because planning makes the robot clean only one room repeatedly.

D Because planning allows the robot to avoid cleaning altogether.

💡 Hint

Think about how planning helps save time and effort.

Planning helps break down a big task into smaller steps and decide the best order to do them. This saves time and avoids repeating work.

❓ Model Choice

intermediate

Choosing a model for planning complex tasks

You want to build an AI that plans a route for a delivery drone to visit many locations. Which model type is best for planning this complex task?

A A simple linear regression model predicting delivery times.

B A clustering model grouping delivery locations by similarity.

C A reinforcement learning model that learns to plan routes by trial and error.

D A basic feedforward neural network classifying package types.

💡 Hint

Think about which model learns sequences of actions to reach a goal.

Reinforcement learning models learn to make sequences of decisions, which is essential for planning routes in complex tasks.

❓ Metrics

advanced

Evaluating planning performance in AI

You have an AI agent that plans tasks. Which metric best measures how well the agent plans complex tasks?

A Total reward accumulated by completing planned tasks efficiently.

B Accuracy of classifying images.

C Mean squared error of predicted values.

D Number of layers in the neural network.

💡 Hint

Consider what shows success in completing planned tasks.

Total reward reflects how well the agent completes tasks efficiently, which is key for planning performance.

🔧 Debug

advanced

Debugging a planning failure in an AI agent

An AI agent planning a multi-step task keeps repeating the same step and never finishes. What is the most likely cause?

A The agent's input data is too large.

B The agent has too many layers in its neural network.

C The agent uses supervised learning instead of reinforcement learning.

D The agent's reward function does not penalize repeated steps.

💡 Hint

Think about what controls the agent's behavior in planning.

If the reward function does not discourage repeating steps, the agent may loop endlessly without progress.

❓ Predict Output

expert

Output of a simple planning algorithm code

What is the output of this Python code simulating a simple plan execution?

Agentic AI

plan = ['start', 'move', 'pick', 'move', 'drop', 'end']
executed = []
for step in plan:
    executed.append(step)
    if step == 'move' and executed.count('move') >= 2:
        break
print(executed)

A ['start', 'move', 'pick', 'move']

B ['start', 'move', 'pick', 'move', 'drop', 'end']

C ['start', 'move', 'pick']

D ['start', 'move', 'pick', 'move', 'drop']

💡 Hint

Count how many times 'move' is added before the loop breaks.

The loop stops when 'move' has been added twice, so the list ends after the second 'move'.