Bias detection and fairness metrics in MLOps - Step-by-Step Execution

Process Flow - Bias detection and fairness metrics

Collect Data

↓

Train Model

↓

Evaluate Model Performance

↓

Calculate Fairness Metrics

↓

Detect Bias?

No→Deploy Model

Yes↓

Mitigate Bias

↓

Re-evaluate Metrics

↩Back to Detect Bias?

This flow shows how after training a model, we check fairness metrics to detect bias. If bias is found, we mitigate it and re-check before deployment.

Execution Sample

MLOps

from sklearn.metrics import accuracy_score, confusion_matrix
# Calculate fairness metric: demographic parity difference
def demographic_parity_difference(y_true, y_pred, sensitive_attr):
    # Calculate positive prediction rates per group
    pass  # simplified for trace

This code snippet calculates a fairness metric called demographic parity difference to detect bias in model predictions.

Process Table

Step	Action	Group A Positive Rate	Group B Positive Rate	Demographic Parity Difference	Bias Detected
1	Calculate positive rate for Group A	0.7	-	-	-
2	Calculate positive rate for Group B	-	0.4	-	-
3	Calculate difference	-	-	0.3	-
4	Compare difference to threshold 0.1	-	-	0.3	Yes
5	Bias mitigation applied	-	-	-	-
6	Recalculate positive rates after mitigation	0.6	0.55	0.05	No
7	Bias check passed	-	-	0.05	No

💡 Bias difference below threshold after mitigation, model ready for deployment

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 6	Final
Group A Positive Rate	undefined	0.7	0.7	0.7	0.6	0.6
Group B Positive Rate	undefined	undefined	0.4	0.4	0.55	0.55
Demographic Parity Difference	undefined	undefined	undefined	0.3	0.05	0.05
Bias Detected	undefined	undefined	undefined	Yes	No	No

Key Moments - 3 Insights

Why do we calculate positive prediction rates separately for each group?

Because bias is about differences in treatment between groups, so we must measure how often each group gets positive predictions separately (see execution_table steps 1 and 2).

What does the demographic parity difference tell us?

It shows how much the positive prediction rates differ between groups. A large difference means bias (see step 3 and 4 in execution_table).

Why do we re-calculate metrics after bias mitigation?

To check if the mitigation worked and bias is reduced below the acceptable threshold (see steps 6 and 7).

Visual Quiz - 3 Questions

Look at the execution table, what is the demographic parity difference at step 3?

A 0.7

B 0.3

C 0.4

D 0.1

💡 Hint

Check the 'Demographic Parity Difference' column at step 3 in the execution_table.

At which step does the bias detection indicate 'Yes'?

A Step 2

B Step 6

C Step 4

D Step 7

💡 Hint

Look at the 'Bias Detected' column in the execution_table to find when bias is first detected.

If the positive rate for Group B after mitigation was 0.3 instead of 0.55, what would happen to bias detection at step 7?

A Bias would still be detected (Yes)

B Bias would not be detected (No)

C Bias detection would be inconclusive

D Bias detection would be skipped

💡 Hint

Compare the difference between Group A and B positive rates after mitigation in variable_tracker.

Concept Snapshot

Bias detection checks if model predictions treat groups fairly.
Calculate positive prediction rates per group.
Measure difference with fairness metrics like demographic parity difference.
If difference > threshold, bias is detected.
Apply mitigation and re-check metrics before deployment.

Full Transcript

Bias detection and fairness metrics help us find if a machine learning model treats different groups fairly. We first collect data and train the model. Then, we calculate how often the model predicts positive outcomes for each group separately. We find the difference between these rates, called demographic parity difference. If this difference is bigger than a set limit, we say the model is biased. We then apply methods to reduce bias and check again. If bias is low enough, the model is ready to deploy. This process ensures fairness in automated decisions.

Practice

1. What is the main purpose of bias detection in machine learning models?

easy

A. To improve the speed of model training

B. To find unfair treatment or discrimination in model predictions

C. To increase the size of the training dataset

D. To reduce the cost of cloud computing resources

Solution

1. Step 1: Understand bias detection context

   Bias detection focuses on identifying unfair or unequal treatment of different groups by a model.

2. Step 2: Compare options to purpose

   Only To find unfair treatment or discrimination in model predictions correctly describes bias detection as finding unfair treatment in predictions.

3. Final Answer:

   To find unfair treatment or discrimination in model predictions -> Option B

4. Quick Check:

   Bias detection = find unfair treatment [OK]

Hint: Bias detection finds unfairness in model results [OK]

Common Mistakes:

• Confusing bias detection with model speed optimization
• Thinking bias detection changes dataset size
• Mixing bias detection with cost reduction

2. Which of the following is a correct way to calculate demographic parity difference in Python?

easy

A. dp_diff = abs(rate_group1 - rate_group2)

B. dp_diff = rate_group1 + rate_group2

C. dp_diff = rate_group1 * rate_group2

D. dp_diff = rate_group1 / rate_group2

Solution

1. Step 1: Understand demographic parity difference formula

   It is the absolute difference between positive outcome rates of two groups.

2. Step 2: Match formula to options

   dp_diff = abs(rate_group1 - rate_group2) correctly uses absolute difference, others use incorrect operations.

3. Final Answer:

   dp_diff = abs(rate_group1 - rate_group2) -> Option A

4. Quick Check:

   Demographic parity difference = absolute difference [OK]

Hint: Use absolute difference for parity difference [OK]

Common Mistakes:

• Using addition or multiplication instead of difference
• Forgetting to take absolute value
• Dividing rates which is not standard

3. Given the following Python code snippet, what is the output?

group1_positive_rate = 0.7
group2_positive_rate = 0.5
dp_diff = abs(group1_positive_rate - group2_positive_rate)
print(round(dp_diff, 2))

medium

A. 0.2

B. 1.20

C. 0.12

D. 0.35

Solution

1. Step 1: Calculate difference between rates

   0.7 - 0.5 = 0.2

2. Step 2: Apply absolute and rounding

   Absolute value is 0.2, rounded to 2 decimals is 0.2

3. Final Answer:

   0.2 -> Option A

4. Quick Check:

   abs(0.7 - 0.5) = 0.2 [OK]

Hint: Subtract and round absolute difference [OK]

Common Mistakes:

• Mixing up subtraction order
• Not rounding output
• Confusing decimal places

4. You wrote this code to calculate equal opportunity difference but it gives wrong results:

tpr_group1 = 0.8
tpr_group2 = 0.6
equal_opp_diff = tpr_group1 - tpr_group2
print(equal_opp_diff)

What is the likely issue?

medium

A. You need to add the true positive rates, not subtract

B. You should multiply the true positive rates instead of subtracting

C. You forgot to take the absolute value of the difference

D. The variable names are incorrect

Solution

1. Step 1: Understand equal opportunity difference metric

   It measures the absolute difference between true positive rates of groups.

2. Step 2: Check code calculation

   Code subtracts but does not take absolute value, so negative results possible.

3. Final Answer:

   You forgot to take the absolute value of the difference -> Option C

4. Quick Check:

   Equal opportunity difference = absolute difference [OK]

Hint: Always use absolute difference for fairness metrics [OK]

Common Mistakes:

• Ignoring absolute value leads to negative results
• Using addition or multiplication wrongly
• Assuming variable names cause errors

5. You want to ensure fairness in a loan approval model. The model predicts positive outcomes for two groups with rates 0.65 and 0.55. Which fairness metric and threshold would best detect bias if you want less than 10% difference between groups?

hard

A. Use accuracy score and check if difference < 0.1

B. Use recall difference and check if difference > 0.2

C. Use precision difference and check if difference > 0.1

D. Use demographic parity difference and check if difference < 0.1

Solution

1. Step 1: Identify appropriate fairness metric

   Demographic parity difference measures difference in positive prediction rates between groups.

2. Step 2: Apply threshold for bias detection

   Checking if difference is less than 0.1 (10%) ensures fairness within acceptable limits.

3. Final Answer:

   Use demographic parity difference and check if difference < 0.1 -> Option D

4. Quick Check:

   Demographic parity difference < 0.1 = fairness [OK]

Hint: Demographic parity difference < threshold detects bias [OK]

Common Mistakes:

• Using accuracy instead of fairness metrics
• Checking for difference greater than threshold incorrectly
• Confusing precision or recall with demographic parity