AI0-001 AI Concepts and Techniques — 20 Questions

Question 1

A data scientist is building a model to predict whether a loan application will default. The dataset has 10,000 labeled examples with 1,000 defaults. Which metric is MOST appropriate for evaluating this highly imbalanced binary classification?

Accepted Answer

AUC-ROC. AUC-ROC is robust to class imbalance because it measures the trade-off between true positive rate and false positive rate across all thresholds. Accuracy is misleading when classes are imbalanced. Precision and recall focus on one class but are threshold-dependent.

Answer

Precision

Answer

Recall

Answer

Accuracy

Question 2

A company is deploying a large language model for customer support. They want to reduce the number of off-topic or nonsensical responses while maintaining creativity. Which parameter adjustment would BEST achieve this?

Accepted Answer

Decrease temperature to 0.2. Lowering temperature makes the model more deterministic and less likely to produce random outputs. Top-p and top-k can also help but are secondary; temperature directly controls randomness.

Answer

Set top-p to 0.1

Answer

Increase top-k to 100

Answer

Increase temperature to 0.9

Question 3

A startup wants to identify unusual patterns in network traffic to detect potential security breaches. They have a large dataset of normal traffic but very few labeled attacks. Which machine learning approach is MOST suitable?

Accepted Answer

Unsupervised anomaly detection. Unsupervised anomaly detection is the most suitable approach because the startup has a large dataset of normal traffic but very few labeled attacks. This technique learns the baseline of normal behavior from unlabeled data and flags deviations as potential anomalies, which is ideal for detecting unknown or rare attack patterns without requiring labeled attack samples.

Answer

Supervised classification with logistic regression

Answer

Reinforcement learning

Answer

Semi-supervised learning

Question 4

A research team is training a deep learning model for image classification using a small dataset of 1,000 labeled images. They are concerned about overfitting. Which combination of regularisation techniques would be MOST effective?

Accepted Answer

Dropout with a rate of 0.5 and L2 regularisation. Dropout randomly disables neurons during training to prevent co-adaptation, and L2 regularisation penalises large weights. Both are standard regularisation techniques. L1 promotes sparsity but is less common for dense layers. Batch normalisation helps convergence but is not primarily a regularisation method.

Answer

Use early stopping without any other regularisation

Answer

L1 regularisation and batch normalisation

Answer

Increase learning rate and use momentum

Question 5

A developer is building a natural language processing system to classify customer reviews as positive, neutral, or negative. They have 50,000 labeled reviews. Which model architecture is MOST appropriate for this task?

Accepted Answer

Fine-tune a pre-trained BERT model. Fine-tuning a pre-trained BERT model is most appropriate because BERT is a transformer-based model pre-trained on a large corpus and can be fine-tuned on the 50,000 labeled reviews to achieve high accuracy with relatively little data. It captures bidirectional context, which is crucial for sentiment classification, and avoids the need for training from scratch.

Answer

Use a convolutional neural network (CNN) on raw text

Answer

Train a recurrent neural network (RNN) from scratch

Answer

Word2vec embeddings followed by logistic regression

Question 6

A company wants to recommend products to users based on their past purchase history. Which machine learning paradigm is BEST suited for this task?

Accepted Answer

Supervised learning with regression. Recommender systems are a classic application of supervised learning (if using regression or classification to predict ratings) or unsupervised learning (collaborative filtering). Among the options, supervised learning with regression is appropriate for predicting purchase likelihood.

Answer

Reinforcement learning

Answer

Unsupervised clustering

Answer

Self-supervised learning

Question 7

A machine learning engineer is training a logistic regression model and notices that the loss is decreasing very slowly. The learning rate is set to 0.001. What is the MOST likely cause and appropriate fix?

Accepted Answer

The learning rate is too low; increase it to 0.01. A learning rate of 0.001 is very low for many logistic regression implementations, causing the gradient descent algorithm to take extremely small steps toward the minimum of the loss function. This results in a slow decrease in loss because each weight update is minimal. Increasing the learning rate to 0.01 allows larger steps per iteration, accelerating convergence without typically causing divergence in well-scaled data.

Answer

The learning rate is too high; decrease it to 0.0001

Answer

The model is overfitting; add L2 regularisation

Answer

The batch size is too large; reduce it

Question 8

A team is training a generative adversarial network (GAN) to generate realistic images of furniture. The generator loss decreases sharply while the discriminator loss increases. What is the MOST likely issue and recommended action?

Accepted Answer

The generator is too strong; train the discriminator more frequently. If the generator loss drops too fast and discriminator loss rises, the generator is overpowering the discriminator. The typical remedy is to train the discriminator more often (e.g., 5 steps per generator step) or adjust the learning rates.

Answer

Mode collapse has occurred; increase the generator's learning rate

Answer

The discriminator is overfitting; decrease its capacity

Answer

The learning rates are too high; reduce both

Question 9

An AI practitioner needs to extract key phrases from a large collection of customer support emails for trend analysis. Which technique is MOST suitable?

Accepted Answer

Named entity recognition (NER). Named entity recognition (NER) identifies predefined entities like names, dates, or product names. For extracting key phrases, keyword extraction or topic modeling (e.g., TF-IDF, RAKE) is more appropriate. Sentiment analysis gives sentiment scores, and text classification assigns categories.

Answer

Language translation

Answer

Text classification

Answer

Sentiment analysis

Question 10

A data scientist is selecting a model for a binary classification task where interpretability is critical because of regulatory requirements. The dataset has 20 features and 10,000 samples. Which model is MOST appropriate?

Accepted Answer

Decision tree. Decision trees are inherently interpretable, showing the decision rules. Random forests and gradient boosting are ensembles that sacrifice interpretability for accuracy. Neural networks are black-box models.

Answer

Neural network (MLP)

Answer

Gradient boosting machine

Answer

Random forest classifier

Question 11

A developer is using a pre-trained BERT model for a question-answering system. They want to ensure the model can handle out-of-vocabulary words. Which component of the BERT architecture is responsible for this?

Accepted Answer

WordPiece tokenisation. WordPiece tokenisation is the component of BERT that handles out-of-vocabulary (OOV) words by breaking them into subword units (e.g., 'playing' → 'play' + '##ing'). This allows the model to represent any word, even unseen ones, as a sequence of known subword tokens, ensuring no word is truly out of vocabulary.

Answer

Positional encoding

Answer

Feed-forward layers

Answer

Attention mechanism

Question 12

A team is training a recurrent neural network (RNN) with LSTM units to predict stock prices. The validation loss is significantly higher than the training loss. Which action is MOST likely to reduce the gap?

Accepted Answer

Increase the dropout rate in LSTM layers. A large gap between training and validation loss indicates overfitting. Increasing dropout (a regularisation technique) reduces overfitting by preventing co-adaptation of neurons. Increasing LSTM units or epochs would worsen overfitting, and reducing sequence length may lose important temporal patterns.

Answer

Increase the number of LSTM units

Answer

Increase the number of training epochs

Answer

Reduce the sequence length

Question 13

A company is deploying a chatbot using a large language model. They want to mitigate the risk of prompt injection attacks. Which TWO measures should be implemented?

Accepted Answer

Implement input validation and sanitisation. Input validation and sanitisation (A) prevent malicious user inputs from being interpreted as instructions by the LLM, directly mitigating prompt injection by stripping or escaping special characters and control sequences. A strict system prompt (B) defines the chatbot's role and boundaries, reducing the attack surface by making it harder for injected prompts to override the intended behavior.

Answer

Fine-tune the model on safe conversational examples

Answer

Use a larger context window

Answer

Limit the maximum output token length

Question 14

A data scientist is evaluating a binary classifier for a medical diagnosis task. The dataset is imbalanced with 5% positive cases. Which THREE metrics should the data scientist consider for a comprehensive evaluation?

Accepted Answer

Precision. Precision (A) is correct because it measures the proportion of true positive predictions among all positive predictions, which is critical in imbalanced medical diagnosis where false positives can lead to unnecessary stress or procedures. In a dataset with only 5% positive cases, a model that predicts all negatives would achieve high accuracy but zero precision, so precision helps assess the cost of false alarms.

Answer

Accuracy

Answer

Perplexity

Question 15

A company wants to build a system that can generate new product images for an online catalog. Which TWO generative AI approaches are most suitable?

Accepted Answer

Diffusion models. Generative Adversarial Networks (GANs) are widely used for image generation, and diffusion models (like Stable Diffusion) have achieved state-of-the-art results in image synthesis. Variational autoencoders (VAEs) can generate images but often produce blurrier outputs. GPT is for text, and BERT is for understanding.

Answer

Variational autoencoders (VAEs)

Answer

BERT-based model

Answer

GPT-style language model

Question 16

A company wants to build a customer service chatbot that answers questions about their internal policy documents. The documents are updated monthly, and the team cannot afford to retrain a model each time. Which approach is MOST appropriate?

Accepted Answer

Use Retrieval-Augmented Generation (RAG) with the policy documents indexed in a vector store. Retrieval-Augmented Generation (RAG) is the most appropriate approach because it allows the chatbot to answer questions based on the latest policy documents without retraining the model. By indexing the documents in a vector store and retrieving relevant chunks at query time, RAG provides up-to-date, contextually accurate answers while keeping the underlying LLM static, which avoids the cost and complexity of monthly retraining.

Answer

Use a larger foundation model with a longer context window and paste all documents into each prompt

Answer

Fine-tune a base LLM on the policy documents monthly

Answer

Train a custom model from scratch on the policy documents each month

Question 17

A data scientist is building a model to predict whether a credit card transaction is fraudulent, using labeled historical data. Which machine learning paradigm is being used?

Accepted Answer

Supervised learning. Supervised learning uses labeled data to train a model to map inputs to outputs. Fraud detection with historical labels is a classic binary classification problem.

Answer

Reinforcement learning

Answer

Unsupervised learning

Answer

Self-supervised learning

Question 18

A deep learning engineer is training a transformer model and notices that validation perplexity increases after a few epochs while training perplexity continues to decrease. Which of the following is the MOST likely cause?

Accepted Answer

The model is overfitting the training data. Option D is correct because the described pattern—decreasing training perplexity alongside increasing validation perplexity—is the classic signature of overfitting. The model is memorizing the training data rather than learning generalizable patterns, causing its performance on unseen validation data to degrade after a certain point in training.

Answer

The temperature parameter is set too high

Answer

The batch size is too small

Answer

The learning rate is too low

Question 19

A product team wants a system that can generate high-quality synthetic images of furniture in different room settings for an online catalog. The images must be photorealistic and vary in style. Which generative AI approach is BEST suited for this task?

Accepted Answer

Diffusion model. Diffusion models are the best choice because they iteratively denoise random noise to produce high-quality, photorealistic images with diverse styles. Unlike GANs, they avoid mode collapse and training instability, and they generate more detailed and varied outputs than VAEs, making them ideal for furniture catalog images in different room settings.

Answer

Variational autoencoder (VAE)

Answer

Generative adversarial network (GAN)

Answer

Recurrent neural network (RNN)

Question 20

An AI practitioner needs to measure the performance of a binary classification model for disease detection, where the cost of false negatives is very high. Which metric should be prioritized?

Accepted Answer

Recall. Recall (true positive rate) minimises false negatives, which is critical when missing a positive case is dangerous.

Answer

Precision

Answer

F1-score

Answer

Accuracy