Quiz: Fundamentals of Artificial Intelligence and Machine Learning — 8 domande

1. What is Artificial Intelligence (AI) primarily considered to be?

A. The study of agents that perceive their environment and take actions to maximize their chances of success.

B. A branch of computer science focused solely on data storage and retrieval.

C. The development of computer hardware capable of running complex algorithms.

D. The simulation of human intelligence processes by machines.

Answer: (D)

Spiegazione

The correct answer is that AI is the simulation of human intelligence processes by machines, as defined by Russell and Norvig. The other options are plausible but do not accurately define AI: the first describes an agent-based perspective, the second relates to hardware, and the fourth pertains to data management, not AI.

2. Who is credited with defining Machine Learning as systems that improve from data without being explicitly programmed, in 1959?

A. Yann LeCun

B. Andrew Ng

C. Arthur Samuel

D. Geoffrey Hinton

Answer: (C)

Spiegazione

Arthur Samuel is credited with coining the term 'Machine Learning' in 1959 and defining it as systems that learn from data and improve their performance without explicit programming.

3. What is the primary role of supervised learning in machine learning?

A. To automatically extract features from raw data without labels

B. To learn from labeled data to make predictions or classifications

C. To optimize decision-making through trial and error in an environment

D. To discover hidden patterns in unlabeled data

Answer: (B)

Spiegazione

Supervised learning's main role is to learn from labeled data in order to make accurate predictions or classifications, as it uses input-output pairs to train models.

4. When was the K-Means clustering algorithm, a fundamental technique in Unsupervised Learning, introduced?

A. 1986

B. 1959

C. 1967

D. 1998

Answer: (C)

Spiegazione

The K-Means clustering algorithm was introduced by MacQueen in 1967, making it a key milestone in the development of Unsupervised Learning techniques.

5. How are Deep Learning and Neural Networks different or similar?

A. Neural Networks are a subset of Deep Learning models.

B. Deep Learning models are a type of Neural Network with multiple layers.

C. Deep Learning is a broader concept that encompasses Neural Networks.

D. Deep Learning and Neural Networks are completely unrelated.

Answer: (B)

Spiegazione

Deep Learning is a subset of Neural Networks characterized by multiple layers, enabling more complex data representations. Therefore, they are similar concepts, with Deep Learning being a specialized form of Neural Networks.

6. Who formulated the first computational model of a neuron, laying the foundation for neural networks?

A. Nair & Hinton

B. Rumelhart et al.

C. McCulloch and Pitts

D. LeCun et al.

Answer: (C)

Spiegazione

McCulloch and Pitts are credited with creating the first computational model of a neuron in 1943, which laid the groundwork for neural networks.

7. What is a primary consequence of effective model evaluation in machine learning?

A. It guarantees the model will perform perfectly on all data

B. It eliminates the need for further model tuning or validation

C. It ensures the model is free from any bias or errors

D. It helps identify whether the model generalizes well to unseen data

Answer: (D)

Spiegazione

Effective model evaluation helps determine whether a machine learning model generalizes well to unseen data, which is crucial for deploying reliable and robust models in real-world applications.

8. How is supervised learning typically applied in practice?

A. Training models on labeled datasets to predict outcomes or classify data

B. Using reinforcement signals to improve the model's decision-making over time

C. Applying algorithms to data without prior knowledge of labels or categories

D. Using unlabeled data to discover hidden patterns in the data

Answer: (A)

Spiegazione

Supervised learning is applied by training models on labeled datasets, where each input has a known output, to enable the model to predict or classify new data accurately.