Learn about robotics.
- So far you've seen how artificial intelligence can work in a world of data. Artificial neural networks can recognize cats, classify music, and make complex connections. But not all AI technology is limited to the digital world. It might be true that a personal assistant like Siri spends most of their time crunching data, but at the end of the day you're not really interested in how well they did their homework. They still need to speak, understand, and respond.
One of the best ways to connect up with humans is to join us in the physical world. That's why one of the most interesting areas in artificial intelligence is the work around robotics. Robotics is about having machines work on physical tasks. It's been around for thousands of years. Inventors have long been fascinated with finding ways to make machines behave like living objects. Up until recently, robotics has been limited to creating highly specialized machines.
The auto plants near my hometown employed several specialized robots. They could do things that even the strongest humans would find impossible. Some could easily lift a car and install parts underneath. As impressive as they were, these robots were very limited in what they could accomplish. Unless they were reprogrammed, they couldn't help a coworker open a car door or start painting the hood. They worked best for repetitive tasks. Combining robotics with artificial intelligence has the potential to give you many more options.
The machine could adapt to its environment and learn many new tasks on the job. A basic example of this is an intelligent house cleaning robot like the Roomba. It uses a form of planning AI to create a map of your different rooms. Then it goes through and vacuums while you're at work. It's careful to avoid stairs and other obstacles that might lead you to come home to its smoldering remains. In many ways, the Roomba is similar to those old industrial machines.
Its programming was hard-coded by an expert. Then there are only a few rules to keep it from tumbling down the stairs. In fact, many need an electronic gate to keep it from tumbling in the same place. A more complex example is the work around autonomous vehicles. The newest vehicles are using supervised machine learning on an artificial neural network. These vehicles are outfitted with many different types of complex sensors. They feed data into the network and the machine looks for patterns for successful driving.
That's why you often see these cars with a human in the driver's seat. They let the neural network process data from the outside world. When it makes an error, the human being will step in and make a correction. Then it will typically use a form of back propagation to help the car learn from its mistakes. Because autonomous cars are a massive project, not many organizations are working on these vehicles. There are many projects around other applications in artificial intelligence and robotics.
One of the key challenges is figuring out what type of artificial intelligence might work best with your robot. Say you wanted to create a robot that distributes prescription medication for seniors in a retirement community. You wouldn't necessarily want to use the same technology they use for autonomous vehicles. Your robots won't run into rain, snow, or trees in the shiny corridors of a retirement community. For this robot you might not want to use machine learning.
There just isn't as much data to analyze. On top of this, there's a huge cost to making errors. It could be dangerous if you gave the wrong prescription medicine to several senior citizens. There might be a lot more advantages to using planning AI for this type of robot. It might be tedious to have an expert program every second of the robot's day. But it potentially has a much smaller chance of making an error. You don't want this robot to learn from its mistakes.
One of the key things to figure out is the role data plays with the robot. Google famously said that they think of their self-driving cars not as a robotics problem but as a big data problem. It's true that to figure out how to get a car to steer left or right is trivial compared to having a car understand when to turn left or right. These robotic cars will need this level of understanding to safely drive on the road.
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