Listen&Learn: Machine learning

Learn how an AI updates its own programming

Pre-listening vocabulary

• artificial intelligence: a program that is built to “think” like a human
• label: to give something a name
• neuron: a cell that sends information through the brain and body
• variation: a change, especially an unexpected one
• fraud: a crime where a person lies or acts unfairly to gain money
• diagnose: to detect a problem, especially a medical one

Listening activity

Podcast: Play in new window | Download (Duration: 1:17 — 1.2MB)

Subscribe: Apple Podcasts | Google Podcasts | RSS | More

Gapfill exercise

Machine learning is the process that artificial intelligence uses to understand . A program using machine learning can update its own programming. A human first trains the AI using a selection of , and the AI then uses its training to recognize in different data. Machine learning can be supervised or unsupervised. In supervised learning, the programmer labels the data so the AI knows what to look for. In unsupervised learning, the AI identifies any patterns that it detects. A complex type of machine learning, called a neural network, is designed almost like a human , with interconnected artificial neurons. These networks are useful for finding complicated patterns that have exceptions and variations. They can detect fraud, make stock predictions, and diagnose issues.

Comprehension questions

See answers below

1. An AI using machine learning
   a. can reject its own programming
   b. can destroy its own programming
   c. can update its own programming
   
2. In supervised machine learning,
   a. the programmer labels data so the AI knows what to look for
   b. the AI identifies any patterns it detects in the unlabelled data
   c. the AI can’t change its programming unless the programmer allows it to
   
3. The type of machine learning that works almost like a brain is called
   a. an artificial neuron
   b. a neural network
   c. an unsupervised detector

Discussion/essay questions

1. The use of artificial intelligence in many industries is controversial. Some people believe it will replace human intelligence and creativity. Do you think it is possible for a machine to think and create the way a human does? What should we do if that happens?

Transcript

Machine learning is the process that artificial intelligence uses to understand information. A program using machine learning can update its own programming. A human first trains the AI using a selection of data, and the AI then uses its training to recognize patterns in different data. Machine learning can be supervised or unsupervised. In supervised learning, the programmer labels the data so the AI knows what to look for. In unsupervised learning, the AI identifies any patterns that it detects. A complex type of machine learning, called a neural network, is designed almost like a human brain, with interconnected artificial neurons. These networks are useful for finding complicated patterns that have exceptions and variations. They can detect fraud, make stock predictions, and diagnose health issues.

Answers to comprehension questions

1c 2a 3b