Learn about data science.
- Data science is another key area where you'll see a lot of crossover with artificial intelligence. In fact, there are many vocal data scientists who also specialize in AI. This makes a lot of sense because many of the active areas in AI are around machine learning and large data sets. Data science is a little bit tricky to define. The most popular definitions revolve around what a data scientist does. That's why most definitions about data science are about knowledge of programing, data, math, statistics, and hacking.
Then a data scientist applies these skills to explore the data using the scientific method. They experiment with data and see if they could discover new stories and insights. Even with all these similarities, data science is still quite a bit different from artificial intelligence. With AI, you're having the machine do most of your pattern matching. It's not exploring to see what you come up with. Instead, it's about throwing everything into the machine and seeing what comes out.
Think of it this way. Imagine that you're a hospital and you want to look for patterns in how you treated your patients. You could approach this challenge from several different angles. You could decide to create your own hospital data science team. It would have a specialist and they ask the team a few critical questions. They might ask a question like, "What was the most popular treatment?" Then the data science team would look through the different data sets. They come up with reports and discuss their findings.
The whole process would be a tight loop of human interaction. The specialist would ask questions and the team would give them reports. Then those reports would probably lead to more questions. Another approach would be to use machine learning. Then you could throw as much data as you can into an artificial neural network hoping to see patterns. Remember that a neural network is like a black box. The network doesn't tell you how it found these patterns. It simply tells you that they exist.
Each of these approaches has its own pros and cons. The data science team will probably know much more about the data behind their insights. They get an intuitive feel for the data and start to ask more interesting questions. The neural network is more likely to find unexpected patterns. The big downside to the neural network is it can't explain to you why these patterns exist. Let's say that the hospital finds that one type of antibiotic has a much higher success rate.
The hospital could easily change their policies to favor the more successful antibiotic, but they won't know why it's more successful. The neural network found a pattern, but it's a black box. They can't ask the network to explain to humans why the pattern exists. On the other hand, the data science team will have a much better feel for their own data. They would have asked questions and use human interaction to find key insights. They wouldn't have looked at all the possibilities like the neural network.
Instead, they would have zoomed right in and asked the crucial question, "Do some of our antibiotics work better than others?" So in a sense, the neural network simply finds the pattern that the antibiotics work better. It's not really interested in why they work better. The data science team should explore why they work better in the search for more data and insights. If you're trying to decide between data science or AI for your project, then you should think of it almost like fishing.
You can use data science to try to fully understand your school of fish. You can ask interesting questions and try to predict their behavior. Maybe you want to know if they prefer warm water or try to understand why they swim together. On the other hand, you can have an artificial neural network take a look at the whole ocean. The network might see patterns, but not necessarily understand cause and effect. It can see the fish schooling together, but it won't explain why. It might be the warmer water or even a migration.
If you don't mind that these patterns are in a black box, then you'll probably get more unexpected insights. If you need to know why you see these patterns, then it might be better to stick with smaller human-driven data science teams.
This course will introduce you to some of the key concepts behind artificial intelligence, including the differences between "strong" and "weak" AI. You'll see how AI has created questions around what it means to be intelligent and how much trust we should put in machines. Instructor Doug Rose explains the different approaches to AI, including machine learning and deep learning, and the practical uses for new AI-enhanced technologies. Plus, learn how to integrate AI with other technology, such as big data, and avoid some common pitfalls associated with programming AI.
- The history of AI
- Machine learning
- Technical approaches to AI
- AI in robotics
- Integrating AI with big data
- Avoiding pitfalls