Lecture 1 (slides)

Question 1

What is the general task of AI?

Answer 1

Predict output given specific features of the inputs

Question 2

How are AI and “regular” statsical modeling similair?

Answer 2

• Input features: independent variables
• Output class: dependent variable
• (in fact, ‘neural networks’ can be seen as a form of logistic regression models)

Question 3

What are the key differrences of AI to stastitical modeling?

Answer 3

• We care about predicting something, not about understanding a (causal) process
• Models are highly complex (and multicollinear) and generally seen as ‘black box’

Question 4

What is deep learning?

Answer 4

Fancy term for machine learning with very large models.

Question 5

On what is deep learning based on?

Answer 5

• Very large neural networks
• Trained on enormous amounts of data, e.g. “all of the internet”
• Using massive computing power, especially of GPUs

Question 6

What are key innovations for deep learning?

Answer 6

• Feature layers find patterns in raw input
• Networks can be (pre-)trained based on unannotated data
• Patterns from (pre-)training are transferred to actual task, and fine-tuned on annotated data

Question 7

What are the Core application area of AI?

Answer 7

Research field of NLP or Computational Linguistics

Question 8

Can the computer understand, generate, or translate text?

Answer 8

Generally

Question 9

What is a recent innovations that is based on ‘BERT’:

Answer 9

Decoder-encoder architecture that builds layers of understanding
trained using tasks such as ‘predict the next word’