Measuring agent accuracy and relevance in Agentic AI - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to calculate the accuracy of an agent's predictions.

Agentic AI

accuracy = correct_predictions / [1]

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

Awrong_predictions

Btotal_agents

Ctotal_predictions

Dcorrect_predictions

Attempts:

3 left

💡 Hint

Common Mistakes

Dividing by wrong_predictions instead of total_predictions.

Using correct_predictions twice in the formula.

Accuracy is the ratio of correct predictions to the total number of predictions made.

2fill in blank

medium

Complete the code to compute the relevance score using cosine similarity.

Agentic AI

relevance_score = cosine_similarity(agent_vector, [1])

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

Aquery_vector

Brandom_vector

Cagent_vector

Dnoise_vector

Attempts:

3 left

💡 Hint

Common Mistakes

Comparing the agent vector to itself instead of the query vector.

Using a random or noise vector which does not relate to the query.

Relevance is measured by comparing the agent's output vector to the query vector using cosine similarity.

3fill in blank

hard

Fix the error in the code that calculates F1 score from precision and recall.

Agentic AI

f1_score = 2 * (precision * recall) / [1]

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

Aprecision + recall

Bprecision - recall

Cprecision * recall

Dprecision / recall

Attempts:

3 left

💡 Hint

Common Mistakes

Using subtraction or multiplication in the denominator instead of addition.

Dividing precision by recall instead of adding them.

The F1 score formula divides twice the product of precision and recall by their sum.

4fill in blank

hard

Fill both blanks to create a dictionary of agent accuracies for agents with accuracy above 0.8.

Agentic AI

accuracies = {agent: [1] for agent, [2] in results.items() if accuracy > 0.8}

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

Aaccuracy

Bscore

Attempts:

3 left

💡 Hint

Common Mistakes

Using a variable name other than accuracy in the loop unpacking.

Not using accuracy as the dictionary value.

The dictionary comprehension extracts accuracy for each agent and filters those with accuracy above 0.8.

5fill in blank

hard

Fill all three blanks to filter agents with relevance above 0.7 and create a summary dictionary.

Agentic AI

summary = {agent: [1] for agent, [2] in agent_results.items() if [3] > 0.7}

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

Arelevance

Daccuracy

Attempts:

3 left

💡 Hint

Common Mistakes

Using accuracy instead of relevance in any blank.

Using different variable names inconsistently.

The dictionary comprehension extracts relevance scores for agents and filters those with relevance above 0.7.