lecture 1: Introduction to AI using Python and PyTorch

What is the definition of Artificial Intelligence (AI)?

AI is the simulation of human intelligence processes by machines, especially computer systems.

What are the two broad types of AI?

Narrow AI (specialized for a specific task) and General AI (capable of performing any intellectual task a human can do).

What is the core component of Narrow AI?

Machine Learning

Why is Python the preferred language for AI development?

Python is free, open-source, and supports major AI libraries like TensorFlow and PyTorch.

What are the main steps in using AI to solve engineering problems?

1. Define the problem.

2. Select an appropriate AI model.

3. Select a training procedure and tweak the model.

4. Validate on unseen data.

What is supervised learning?

It involves training an AI model using labeled data where both inputs and expected outputs are known.

Name a common loss function used for regression.

Mean Squared Error (MSE).

What are decision trees used for?

They are used to classify data based on a series of decision rules.

What is the main benefit of using dropout layers in neural networks?

They help avoid overfitting by randomly turning off neurons during training.

What is the difference between classification and regression?

Classification predicts discrete categories, while regression predicts continuous values.

What are some examples of Python libraries used in AI development?

sklearn (machine learning), matplotlib (plotting), numpy (numerical functions), and torch (neural network training).

What does "overfitting" mean in the context of AI models?

Overfitting occurs when a model learns the training data too well, including noise, resulting in poor performance on new data.

What is a neural network?

A model inspired by the human brain that uses layers of artificial neurons to process inputs and produce outputs.

What is reinforcement learning?

A type of machine learning where an agent learns by trial and error to maximize rewards based on its actions.

What are the advantages of using GPUs in AI training?

GPUs allow efficient optimization and fitting of complex models to large datasets due to their parallel processing capabilities.

What is the significance of cross-entropy loss in classifiers?

Cross-entropy loss is used to measure the performance of a classification model whose output is a probability value between 0 and 1.

What is stochastic gradient descent (SGD)?

A method to optimize AI models by updating parameters using random subsets of data (mini-batches).

What is the role of activation functions in neural networks?

Activation functions introduce non-linearities, allowing the network to learn complex patterns in data.

How can overfitting be mitigated in AI models?

By using simpler models, more data, dropout layers, and validation techniques like cross-validation

What is a "hyperplane" in the context of AI?

A hyperplane is a decision boundary in high-dimensional space used for classification tasks.

What is self-supervised learning?

A machine learning approach that uses data to generate its own labels, often applied in text and image tasks.

What are some challenges associated with AI?

Privacy concerns, data collection ethics, bias, hallucinations, and overfitting.

What is a loss function, and why is it important?

A loss function quantifies the difference between predicted and actual values, guiding the optimization of the model.

What are diffusion models used for in AI?

They are used to generate images by training a model to remove random noise.

What is the difference between parametric and non-parametric models?

Parametric models have a fixed number of parameters, while non-parametric models can adapt to data size and complexity.

What are the key steps in training a neural network?

1. Define the network architecture.

2. Choose a loss function.

3. Optimize weights using gradient descent.

4. Validate using unseen data.

Why is data cleaning important in AI projects?

Clean data ensures accurate model training and avoids errors due to missing or inconsistent inputs.

What does the term "big data" imply in AI?

Big data refers to large datasets that provide comprehensive coverage, reducing the risk of overfitting.

What is the role of transformers in modern AI?

Transformers are advanced architectures that excel in natural language processing and have enabled models like ChatGPT.

What is the main advantage of using JupyterLab for Python-based AI projects?

JupyterLab's notebook format is highly interactive and popular for machine learning tasks, making it easy to test and visualize results.