The computer system we use to help our fellow humans is not perfect.
But, it’s better than nothing, says Paul O’Connor, chief scientist at the UK-based company Alpha Computer Systems.
“We’ve created an amazing new system called Alpha, which enables us to create a machine that’s able to make a decision, even when there’s no way to actually do it,” he says.
The Alpha system uses a machine learning algorithm to create models of the world that are then used to create algorithms to decide which actions to take based on these models.
O’Connors team developed Alpha by applying machine learning techniques to the work of computer scientists to create their own software that would be able to automatically determine what to do based on the model of the environment it was built from.
The software could also determine when the computer needed to stop working to fix a problem, for example.
Alpha uses a different approach to human intelligence compared to other computer systems.
Alpha’s AI uses a computer system to learn from experience, rather than a human.
“Instead of looking at all the examples in the literature, we were looking at what are the common scenarios in which the AI could make a good decision,” O’Connell says.
“That’s what we were able to learn about and then develop the algorithms to do.”
Alpha also uses a model of what’s going on in the environment to determine what actions to perform based on this information.
The system then uses this to determine which of these actions would be best.
Alpha also has a “supervised learning” feature that can be used to learn more about the environment in which a computer is operating.
The algorithm that is used to determine when to stop learning a model depends on a large amount of information that is also used to predict the future.
This information includes what the computer is seeing in the real world.
This helps the algorithm to understand what might happen if the system went back to what it was learning.
If the computer goes back to the same model, the algorithm will not learn anything new, but it will be able learn more from the information in the new model.
“What we’ve done is basically given the AI a whole set of training and learning cycles and then we have a model which we then feed back into the machine learning system,” O,Connor says.
This way, the AI is able to continuously learn new things, whether it’s from what the environment looks like, or from what it knows about itself.
“It’s basically an evolution process for the system, and that’s what makes it a supercomputer,” O said.
“Because the AI can adapt to its environment and what it’s seeing, it can also change it.”
Alpha was developed to help the human race survive in the face of a global pandemic, but O’Conner is optimistic about the future of AI.
“AI is going to change everything,” he said.
This is something that we’re all very familiar with, but I think we’re still in the very early stages.
We’re seeing very different things that we see in the media.
The AI is going have a much bigger impact on how we use technology and how we think about things.
And we have to be prepared for that, because we’re in a new era where we’re dealing with artificial intelligence,” O says.
For more information, read: ‘Alpha’ computer systems help solve world’s most complex problems, researchers say in Science