Responsible AI practices in MLOps - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to log model fairness metrics using a popular MLOps tool.

MLOps

from mlflow import [1]

client = [1].MlflowClient()
client.log_metric(run_id, 'fairness_metric', fairness_value)

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

Amodel

Btracking

CMlflow

Dexperiment

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'Mlflow' instead of 'tracking' causes import errors.

The tracking module provides the MlflowClient class to log metrics such as fairness.

2fill in blank

medium

Complete the code to check for bias in a dataset using the AI Fairness 360 toolkit.

MLOps

from aif360.datasets import [1]

dataset = [1](features, labels)

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

AFairnessDataset

BBiasDataset

CStandardDataset

DMLDataset

Attempts:

3 left

💡 Hint

Common Mistakes

Using non-existent classes like 'BiasDataset' causes import errors.

StandardDataset is the common class to load datasets for bias checking in AI Fairness 360.

3fill in blank

hard

Fix the error in the code that applies a fairness metric to a model's predictions.

MLOps

from aif360.metrics import [1]

metric = [1](dataset_true, dataset_pred, unprivileged_groups, privileged_groups)
fairness_score = metric.mean_difference()

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

AClassificationMetric

BBiasMetric

CFairnessMetric

DPredictionMetric

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong class names causes attribute errors.

ClassificationMetric is the correct class to compute fairness metrics on classification results.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that filters features with values greater than 0.

MLOps

{feature: value for feature, value in features.items() if value [1] 0 and feature [2] 'age'}

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

A>

B<

C==

D!=

Attempts:

3 left

💡 Hint

Common Mistakes

Using '<' instead of '>' causes wrong filtering.

Using '==' instead of '!=' includes 'age' feature.

The code filters features where the value is greater than 0 and the feature name is not 'age'.

5fill in blank

hard

Fill all three blanks to create a dictionary comprehension that maps uppercased feature names to values greater than 10.

MLOps

{ [1]: [2] for [3], value in features.items() if value > 10 }

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

Afeature.upper()

Bvalue

Cfeature

Dval

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'val' instead of 'value' causes NameError.

Not uppercasing feature names changes keys.

The comprehension maps uppercased feature names to their values for values greater than 10.