Blog

In order for robots to effectively partake in search and rescue operations, they need to effectively navigate obstacles in their way. One area that is particularly common and difficult to venture into is vegetation.

Robots typically use a combination of sensors to perceive their surroundings such as ultrasonic sensors, Lidar (Light Detection and Ranging), infrared sensors and camera systems. However, these are not often enough to allow robots to actually bypass the vegetation so commonly found in real outdoor environments.

That’s why engineers at Carnegie Mellon University are working on solving this particular dilemma.

Instead of looking at individual neurons, they look at combinations of neurons that form patterns or features.

Artificial neural networks are like digital versions of our brains. They learn from data, not rules, and they can perform extraordinary tasks, from translating languages to playing chess. But how do they do it? What is the logic behind their calculations? And how can we trust them to be safe and reliable?

XH4D/iStock.

AI brains: How do they work?

Understanding causality can’t come from passive observation, because the relevant counterfactuals often do not arise. If X is followed by Y, no matter how regularly, the only way to really know that is a causal relation is to intervene in the system: to prevent X and see if Y still happens. The hypothesis has to be tested. Causal knowledge thus comes from causal intervention in the world. What we see as intelligent behavior is the payoff for that hard work.

The implication is that artificial general intelligence will not arise in systems that only passively receive data. They need to be able to act back on the world and see how those data change in response. Such systems may thus have to be embodied in some way: either in physical robotics or in software entities that can act in simulated environments.

Artificial general intelligence may have to be earned through the exercise of agency.