Why agents make autonomous decisions in Prompt Engineering / GenAI - Challenge Your Understanding

Challenge - 5 Problems

🧠 Conceptual

intermediate

Why do autonomous agents need to make decisions without human input?

Imagine a robot vacuum cleaner working in your home. Why does it need to decide on its own where to clean next instead of waiting for you to tell it?

A Because it cannot communicate with humans at all.

B Because it wants to avoid humans and work secretly.

C Because it is programmed to ignore human commands.

D Because it can respond faster to changes in the environment without waiting for instructions.

💡 Hint

Think about how waiting for human instructions might slow down the robot's work.

Autonomous agents make decisions on their own to react quickly and efficiently to their surroundings without delays caused by waiting for human input.

❓ Model Choice

intermediate

Choosing the right model for autonomous decision-making

You want to build an agent that can decide the best route to deliver packages in a city with changing traffic. Which model type is best suited for this task?

A A reinforcement learning model that learns from trial and error in the environment.

B A clustering model that groups delivery locations.

C A simple rule-based system with fixed instructions.

D A supervised learning model trained on fixed routes only.

💡 Hint

Consider which model can adapt by learning from experience in a changing environment.

Reinforcement learning models are designed to learn optimal actions through interaction with the environment, making them ideal for dynamic decision-making tasks.

❓ Metrics

advanced

Evaluating autonomous agent decision quality

An autonomous agent is tested on how well it completes tasks without human help. Which metric best measures how often it makes the correct decision?

A Latency - the time taken to make a decision.

B Loss - the error between predicted and actual outcomes.

C Accuracy - the percentage of correct decisions made by the agent.

D Throughput - the number of tasks completed per minute.

💡 Hint

Think about which metric directly shows how often the agent's decisions are right.

Accuracy measures the proportion of correct decisions, directly reflecting decision quality in classification or choice tasks.

🔧 Debug

advanced

Why does this autonomous agent fail to learn?

Consider this simplified reinforcement learning code snippet for an agent:

rewards = [1, -1, 1, 1]
actions = ["left", "right", "left", "left"]

for i in range(len(actions)):
    if rewards[i] > 0:
        policy = actions[i]
print(policy)

Why does this code fail to learn the best action?

A Because it overwrites the policy every time instead of accumulating knowledge.

B Because the rewards list has negative values which are not allowed.

C Because the loop does not run due to incorrect range.

D Because the actions list contains strings instead of numbers.

💡 Hint

Look at how the variable policy changes inside the loop.

The code sets the policy to the last positive reward action, losing all previous learning. It should accumulate or update the policy properly.

🧠 Conceptual

expert

Why do autonomous agents balance exploration and exploitation?

In autonomous decision-making, agents often must choose between trying new actions (exploration) and using known good actions (exploitation). Why is this balance important?

A Because only exploring new actions guarantees the best long-term results.

B Because balancing both helps the agent find better options while using what it already knows.

C Because only exploiting known actions avoids any risk of failure.

D Because agents cannot remember past actions, so they randomly choose.

💡 Hint

Think about how trying new things and using what works can both help an agent improve.

Balancing exploration and exploitation allows agents to discover better actions while still benefiting from past knowledge, leading to improved decision-making over time.