Recent artificial intelligence advancements are taking the media by storm. They can perform impressive feats such as:
- object recognition that enables self-driving cars
- playing video games that outperform world best players
- detecting breast cancer faster and more accurate than doctors
These machines can perform tasks and solve problems better than us, but they are only designed to excel in the intended functions. However impressive these may seem to be, we are still decades away from building artificial human intelligence.
Building machines that have truly human-like learning and cognitive that can learn and think as we do has been an ongoing pursuit. What does it mean for a machine to learn or think like a person?
- able to explain and understand problems
- learning-to-learn to acquire knowledge
- generalise knowledge to new tasks and situations
Our understanding of our brain has allowed us to design models which emulating the brain. In a study by KAIST, they developed a computational and neural mechanism for human meta reinforcement learning.
As humans, we can adapt to complexity and uncertainty when learning and making decisions. They aim to build models that can make decisions like humans and solves problems in the same way that humans do.
Their discoveries have opened up the possibility that might lead to machines that can take on more human-like intelligence.
Even with our eyes covered, one may still play the guitar well. But for an artificial intelligence system to play the guitar in a new environment (i.e. with eyes covered), it might not do as well.