Since a computer is limited by its input values and programming, ideas are not something it can deal with.
What is an "idea", beyond an abstract concept? Computers can certainly model concepts.
Or, do you mean "idea" in the sense of inspiration? For that, you just need an ability for reflection, which is to say ways that a network can feed back into itself. Some random event can then trigger associations and if something sensible emerges, that's your idea. Parts of your brain are filtering out nonsensical ideas and interpretations all of the time, even without you being aware of it.
In neural networks, this dynamic has been simulated using an approach called a
generative adversarial network, where two networks feedback on each other. One acts as an idea generator, and the other is a sort of gatekeeper that rejects nonsense. As they contest with each other, each is improved. It's a technique that underlies some of these "AI image generators".
Like asking a computer about a feeling, or emotion.
Define "feeling" or "emotion". I think a lot of people get hung up on these, because we have some intuitive sense of what they are, but we don't think about physiological manifestation or their role in cognition. When you look at them at that level, all that's happening is a chemical change that instantly influences the behavior of the network (quite a gross oversimplification, I know).
I'm quite certain you can emulate moods, in an artificial neural network. The main question is whether you can tune the machinery of how moods influence its function and how moods are regulated as finely as it is in humans. And lets not kid ourselves: we all know people who are
too moody or too quick to anger. So, it's not as if biology has exactly struck a perfect balance.
Facebook
Twitter
Instagram