Lesson 3 of 7 · 12 min
AI and how machine learning learns
AI systems do tasks that once needed human intelligence; machine learning, the key AI tool for Big Data, learns structure from large datasets without assuming a distribution, using separate training, validation and test data.
In short
- Artificial intelligence (AI): computer systems that perform tasks that traditionally required human intelligence, at a level comparable to or better than humans.
- Early AI used expert systems built on if–then rules; banks have used neural networks to flag credit card fraud since the 1980s.
- Machine learning (ML) extracts knowledge from large datasets without assuming a probability distribution; it generalises from known examples to find structure, without human help.
- The data are split into training (find relationships), validation (validate and tune) and test (check prediction on new data) sets.
- ML needs massive, clean data and human judgment; Big Data is what made it practical in investing.
- ML models are not explicitly programmed, so they can be a black box whose results are hard to explain.
Unlock this lesson free for 7 days
Create a free account to get 7 days of full access — every lesson, video, flashcard, mock and the question bank. No card needed.