Introduction#

  • Basic Bias: 21 metrics
  • SageMaker Clarify notebook
  • I calcualte bias notebook

Data Bias Detection#

Some questions about Bias

  • Does the group representation in the training data reflect the real world?
  • Does the model have different accuracy for different groups?
  • Is the model performance equal for different gorups or facets?

Quoted what is bias mean: An imbalance in the training data or the prediction behavior of the model across different groups, such as age or income bracket. Biases can result from the data or algorithm used to train your model. For instance, if an ML model is trained primarily on data from middle-aged individuals, it may be less accurate when making predictions involving younger and older people. Sample Notebook and Blog

How to detecte bias

  • Wrangler bias report
  • Clarify for pre-training and post-training bias analysis

Let use SageMaker Clarify to detect bias in data. Download this bank additional dataset

wget https://sagemaker-sample-data-us-west-2.s3-us-west-2.amazonaws.com/autopilot/direct_marketing/bank-additional.zip

Then unzip and upload to S3

aws s3 cp . s3://bucket-name/bank-additional/ --recursive

Create a wrangler flow, import data, and add an analysis to see the Bias report.

  • Predicted Y: term deposit
  • Bias analysis on column: martial

The Bias report show some basic metrics such as Class Imbalance.

CI = (na - nd)(na + nd)

From the results, we observe a class imbalance where the married class is 21% more represented than other classes. We also observe that the married class is 2.8% less likely to subscribe to a bank term deposit.

SageMaker Clarify - UCI DataSet#

Follow this notebook to understand how to use SageMaker Clarify to detect bias pre-training and post-training.

Let download another dataset which is the UCI dataset. This predict income greater or less than 50K.

from sagemaker.s3 import S3Downloader


adult_columns = [
    "Age",
    "Workclass",
    "fnlwgt",
    "Education",
    "Education-Num",
    "Marital Status",
    "Occupation",
    "Relationship",
    "Ethnic group",
    "Sex",
    "Capital Gain",
    "Capital Loss",
    "Hours per week",
    "Country",
    "Target",
]




S3Downloader.download(
    s3_uri=f"s3://sagemaker-example-files-prod-{region}/datasets/tabular/uci_adult/adult.data",
    local_path="./",
    sagemaker_session=session,
)

Let analyse and visualize it sex column, how male versus female represents in data and in label.

training_data = pd.read_csv(
    "adult.data", names=adult_columns, sep=r"\s*,\s*", engine="python", na_values="?"
).dropna()


training_data["Sex"].value_counts().plot.bar()
plt.title("Count of Sex")
plt.show()


training_data["Sex"].where(training_data["Target"] == ">50K").value_counts().plot.bar()
plt.title("Counts of Sex earning >$50K")
plt.show()

SageMaker Clarify - Prepare Data#

Now let prepare data for training a XGBoost algorithm. We need to convert catalogrical into numerical using a number_encode_features function as below.

from sklearn import preprocessing


def number_encode_features(df: pd.DataFrame):
    """
    convert categorical to numerical
    """
    result = df.copy()
    encoders = {}
    for column in result.columns:
        if result.dtypes[column] == object:
            encoders[column] = preprocessing.LabelEncoder()
            result[column] = encoders[column].fit_transform(
                result[column].fillna("None")
            )


    return result, encoders

Create train and test data

# swap the target as first column
training_data = pd.concat(
    [training_data["Target"], training_data.drop(["Target"], axis=1)], axis=1
)


# swap the target as first column
testing_data = pd.concat(
    [testing_data["Target"], testing_data.drop(["Target"], axis=1)], axis=1
)


# save train data to csv
training_data, _ = number_encode_features(training_data)
training_data.to_csv("train_data.csv", index=False, header=False)


# save test data to csv
testing_data, _ = number_encode_features(testing_data)
testing_data.to_csv("test_features.csv", index=False, header=False)

SageMaker Clarify - Train Model#

Let create TrainingInput object

from sagemaker.inputs import TrainingInput
from sagemaker.s3 import S3Uploader
prefix = "sagemaker-clarify"


# upload train data to s3
train_uri = S3Uploader.upload(
    local_path="./train_data.csv", desired_s3_uri=f"s3://{bucket}/{prefix}"
)


# upload test data to s3
test_uri = S3Uploader.upload(
    local_path="./test_features.csv", desired_s3_uri=f"s3://{bucket}/{prefix}"
)

Let create a XGBoost model

from sagemaker.inputs import TrainingInput
from sagemaker.s3 import S3Uploader


xgboost_image_uri = retrieve("xgboost", region, version="1.5-1")


xgb = Estimator(
    xgboost_image_uri,
    role,
    instance_count=1,
    instance_type="ml.m5.xlarge",
    disable_profiler=True,
    sagemaker_session=session,
)


xgb.set_hyperparameters(
    max_depth=5,
    eta=0.2,
    gamma=4,
    min_child_weight=6,
    subsample=0.8,
    objective="binary:logistic",
    num_round=800,
)


xgb.fit({"train": train_input}, logs=False)

After training, we have to create a Model object

from datetime import datetime
model_name = f"demo-clarify-bias-{datetime.now().strftime('%d-%m-%Y-%H-%M-%S')}"
model = xgb.create_model(name=model_name)
model.create()

SageMaker Clarify - Processor#

To get the bias report pre and post training, we need to run a SageMaker Clarify processing Create SageMaker Clarify Processor

clarify_processor = clarify.SageMakerClarifyProcessor(
    role=role,
    instance_count=1,
    instance_type="ml.m5.xlarge",
    sagemaker_session=session
)

Create Writing DataConfig which communicates some basic information about data I/O to SageMaker Clarify.

data_config = clarify.DataConfig(
    s3_data_input_path=train_uri,
    s3_output_path=bias_report_output_path,
    label="Target",
    headers=training_data.columns.to_list(),
    dataset_type="text/csv",
)

Create Writing ModelConfig which is an object communicates information about your trained model.

model_config = clarify.ModelConfig(
    model_name=model_name,
    instance_type="ml.m5.xlarge",
    instance_count=1,
    accept_type="text/csv",
    content_type="text/csv",
)

Create Writing ModelPredictedLabelConfig provides information on the format of your predictions

prediction_config = clarify.ModelPredictedLabelConfig(probability_threshold=0.8)

Create Writing BiasConfig contains configuration values for detecting bias using a Clarify container.

bias_config = clarify.BiasConfig(
    label_values_or_threshold=[1],
    facet_name="Sex",
    facet_values_or_threshold=[0],
    group_name="Age",
)

Run the Clarify Processor to get bias information of post-training and pre-training

clarify_processor.run_bias(
    data_config=data_config,
    bias_config=bias_config,
    model_config=model_config,
    model_predicted_label_config=prediction_config,
    pre_training_methods="all",
    post_training_methods="all"
)

Finally, you can find the bias report in SageMaker Studio Experiments, or read the report in json format as below

{
  "name": "CI",
  "description": "Class Imbalance (CI)",
  "value": 0.3513692725946555
},
{
  "name": "DPL",
  "description": "Difference in Positive Proportions in Labels (DPL)",
  "value": 0.20015891077100018
},
{
  "name": "JS",
  "description": "Jensen-Shannon Divergence (JS)",
  "value": 0.03075614465977302
},
{
  "name": "KL",
  "description": "Kullback-Liebler Divergence (KL)",
  "value": 0.14306865156306428
},

Reference#