AI0-001 Practice Test 37 — 15 Questions

Question 1

A team is implementing a RAG system for legal document retrieval. The documents are long and cover multiple topics. Which chunking strategy is MOST appropriate to ensure each chunk contains coherent information?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Semantic chunking based on topic boundaries. Semantic chunking based on topic boundaries is the most appropriate strategy because legal documents are long and cover multiple topics. By splitting at natural topic shifts (e.g., clauses, sections, or argument transitions), each chunk preserves coherent meaning, which is critical for accurate retrieval and generation in a RAG system. This approach avoids mixing unrelated content within a single chunk, which would degrade the quality of retrieved context.

Answer

Hierarchical chunking with overlapping windows

Answer

Fixed-size chunking with 512 tokens

Answer

Character-level chunking with no overlap

Question 2

An organisation is deploying a fine-tuned LLM for internal use. They need to ensure the API endpoint is secure and cost-effective. Which TWO measures should they implement? (Choose 2)

Accepted Answer

Implement API key authentication. API key authentication (Option A) is a fundamental security measure that ensures only authorized clients can access the LLM endpoint. It provides a simple, lightweight mechanism to validate requests without the overhead of full OAuth, making it both secure and cost-effective for internal deployments.

Answer

Enable content filtering

Answer

Disable logging to reduce storage costs

Answer

Use gRPC instead of REST

Question 3

Which neural network architecture is specifically designed to process sequential data, such as time series or sentences, by maintaining a hidden state that captures information about previous inputs?

Accepted Answer

Recurrent Neural Network (RNN). Recurrent Neural Networks (RNNs) are specifically designed for sequential data because they maintain a hidden state that is updated at each time step, allowing information about previous inputs to persist and influence current and future outputs. This feedback loop makes them ideal for tasks like time series forecasting, natural language processing, and speech recognition, where order and context matter.

Answer

Transformer

Answer

Convolutional Neural Network (CNN)

Answer

Multi-layer Perceptron (MLP)

Question 4

An MLOps team wants to deploy a trained PyTorch model to production with low latency inference. The model must be interoperable across different frameworks and runtimes. Which approach is BEST?

Accepted Answer

Export the model to ONNX format and deploy using ONNX Runtime. ONNX (Open Neural Network Exchange) provides a standardized, framework-agnostic format that ensures interoperability across different runtimes and hardware accelerators. By exporting the PyTorch model to ONNX and deploying with ONNX Runtime, the team achieves low-latency inference through graph optimizations and hardware-specific execution providers, while avoiding vendor lock-in.

Answer

Deploy the native PyTorch model using TorchServe

Answer

Quantize the model to INT8 and deploy as a TensorFlow Lite model

Answer

Convert the model to TensorFlow SavedModel and deploy using TensorFlow Serving

Question 5

An AI system used for autonomous driving is found to have a lower accuracy in detecting pedestrians with darker skin tones. The development team wants to address this ethical issue. Which action is most effective?

Accepted Answer

Augment the training dataset with more images of pedestrians with darker skin. Augmenting the training dataset with more images of pedestrians with darker skin directly addresses the root cause of the bias: underrepresentation in the training data. By providing a more balanced and diverse dataset, the model can learn more robust features for all skin tones, reducing accuracy disparity without altering the algorithm's core logic or introducing arbitrary thresholds.

Answer

Conduct additional testing to measure the disparity

Answer

Replace the object detection algorithm with a different one

Answer

Adjust the model's decision threshold for pedestrian detection

Question 6

A hospital wants to deploy a machine learning model to predict patient readmission risk within 30 days. They have a dataset with 10,000 records, 70 features including demographics, lab results, and past admissions. The target variable is binary (readmitted or not). The data scientist trains a logistic regression model and achieves an AUC of 0.85 on the test set. However, the hospital's clinicians require interpretability of predictions to trust the model. Which action should the data scientist take to ensure the model meets the interpretability requirement while maintaining performance?

Accepted Answer

Use the logistic regression model as is, since it is inherently interpretable with coefficients. Logistic regression is a linear model whose predictions are computed as a weighted sum of input features passed through a sigmoid function. The coefficients directly represent the log-odds change per unit increase in each feature, making the model inherently interpretable without any post-hoc explanation tools. Since the model already achieves an AUC of 0.85, which meets performance requirements, no architectural change is needed. The data scientist should retain the logistic regression model and present the coefficients (and odds ratios) to clinicians to satisfy the interpretability requirement.

Answer

Reduce the number of features to 10 using PCA and retrain the logistic regression

Answer

Replace logistic regression with a random forest model and use feature importance plots

Answer

Train a deep neural network and apply LIME or SHAP for explanations

Question 7

You are a security engineer at a large e-commerce company that uses an AI-based recommendation system. The system is deployed on a Kubernetes cluster and uses a TensorFlow model served via REST API. Recently, the security team detected unusual API calls that caused the model to return incorrect recommendations. Analysis shows that the inputs were crafted to maximize prediction error. The team suspects an adversarial attack. You need to implement a solution that detects and mitigates such attacks in real-time without requiring model retraining. Which approach should you take?

Accepted Answer

Implement an input validation filter to detect and block anomalous inputs. An input validation filter can detect and block adversarial inputs in real-time by analyzing statistical properties (e.g., outlier detection, perturbation magnitude) without modifying the model. This approach is lightweight, operates at the API gateway level, and does not require retraining, making it suitable for immediate deployment against crafted inputs that maximize prediction error.

Answer

Increase the number of model replicas to distribute the load

Answer

Retrain the model with adversarial examples

Answer

Roll back the model to a previous version that was not attacked

Question 8

A data scientist needs to predict whether a customer will churn based on historical data containing features like account age, monthly charges, and support tickets. The target variable is binary (churn or not). Which type of machine learning algorithm should be used?

Accepted Answer

Logistic regression. Logistic regression is the correct choice because it is specifically designed for binary classification tasks, such as predicting whether a customer will churn (yes/no). It models the probability of the binary outcome using a logistic (sigmoid) function, making it suitable for this supervised learning problem with a categorical target variable.

Answer

Linear regression

Answer

K-means clustering

Answer

Principal component analysis

Question 9

A healthcare AI startup is developing a diagnostic tool that uses patient data to predict disease risk. To comply with HIPAA and minimize privacy risks while still training accurate models, which privacy-preserving technique should they prioritize?

Accepted Answer

Differential privacy. Differential privacy is the correct choice because it provides a formal mathematical guarantee that the output of a model does not reveal whether any individual's data was included in the training set. This is essential for HIPAA compliance as it prevents re-identification attacks even when an adversary has auxiliary information. Unlike anonymization or pseudonymization, differential privacy adds calibrated noise to the training process or query results, ensuring strong privacy protection while preserving model utility.

Answer

Anonymization

Answer

Pseudonymization

Answer

Data minimization

Question 10

A company trains a sentiment analysis model on customer reviews. An attacker submits hundreds of reviews with the word 'excellent' attached to negative feedback, causing the model to classify negative reviews as positive. This is an example of which attack?

Accepted Answer

Data poisoning. Data poisoning occurs when an attacker deliberately corrupts the training data to manipulate the model's behavior. By injecting hundreds of reviews that pair the word 'excellent' with negative sentiment, the attacker shifts the model's learned decision boundary, causing it to misclassify genuinely negative reviews as positive. This directly undermines the integrity of the training dataset, which is the hallmark of a data poisoning attack.

Answer

Model extraction

Answer

Adversarial example

Answer

Prompt injection

Question 11

A data scientist trains a linear regression model to predict house prices. The model has high bias and low variance. Which action would most likely reduce bias?

Accepted Answer

Add polynomial features. High bias indicates the model is underfitting the data, meaning it is too simple to capture the underlying patterns. Adding polynomial features increases model complexity by introducing non-linear terms, which allows the linear regression model to better fit the training data and thus reduce bias.

Answer

Apply L2 regularization

Answer

Increase the training dataset size

Answer

Remove irrelevant features

Question 12

An AI team is developing a model that approves loan applications. The dataset contains historical loan decisions where a protected group was disproportionately denied loans. The team wants to ensure the model does not perpetuate this bias. Which fairness metric should be used during validation to directly measure whether the model's positive prediction rate is equal across groups?

Accepted Answer

Demographic parity. Demographic parity (also called statistical parity) directly measures whether the positive prediction rate is equal across groups. It is the fairness metric that compares the proportion of positive outcomes for each protected group, which aligns with the requirement to measure equal positive prediction rates.

Answer

Calibration

Answer

Individual fairness

Answer

Equalised odds

Question 13

A company uses an AI system to screen job applicants. Under the GDPR, if the system makes automated decisions that have a legal or similarly significant effect on individuals, the data subject has the right to obtain an explanation of the decision. What is this right commonly called?

Accepted Answer

Right to explanation. Under GDPR Article 22, data subjects have the right not to be subject to a decision based solely on automated processing that produces legal or similarly significant effects. Recital 71 and Article 13-15 collectively establish what is commonly called the 'right to explanation' — the right to obtain meaningful information about the logic involved, as well as the significance and envisaged consequences of the automated decision. This is the term used in the question stem.

Answer

Right to erasure (right to be forgotten)

Answer

Right to object

Answer

Right to data portability

Question 14

A healthcare AI system diagnosing diabetic retinopathy from retinal images shows high accuracy overall but significantly lower recall for patients with darker skin tones. Which fairness metric would BEST capture this disparity by comparing true positive rates across groups?

Accepted Answer

Equalised odds. Equalised odds is a fairness metric that requires both true positive rates (TPR) and false positive rates (FPR) to be equal across groups. In this scenario, the disparity is specifically in recall (TPR) for patients with darker skin tones, so equalised odds directly captures the difference in TPR between groups. It ensures that the model's ability to correctly identify positive cases is independent of the group attribute.

Answer

Calibration

Answer

Demographic parity

Answer

Individual fairness

Question 15

A financial institution uses a deep learning model for loan approvals. Under the EU AI Act, this is considered a high-risk AI system. Which mandatory requirement must the institution fulfill before deployment?

Accepted Answer

Conduct a risk assessment and bias testing. Under the EU AI Act, high-risk AI systems must undergo a conformity assessment that includes a risk assessment and bias testing to ensure fairness, transparency, and non-discrimination before deployment. This requirement is mandated by Articles 9 and 10 of the Act, which specifically address risk management and data governance for high-risk systems. Option D correctly identifies this mandatory step, as the institution must demonstrate that the model does not produce biased outcomes that could lead to discriminatory lending practices.

Answer

Obtain certification from an ISO 27001 auditor

Answer

Publish the model's source code publicly

Answer

Register the AI system with the national data protection authority