Researchers at the University of Hertfordshire have developed a new algorithm that allows robots to function more intuitively, using the environment as a guide to make decisions.
The principle is that through algorithms, robot agents create their own goals.
For the first time, the algorithm unifies different goal-setting methods under one concept directly related to physics, and it also makes calculations transparent so that others can study and adopt it.
The principle of this algorithm is related to the famous chaos theory, because this method makes the agent “the master of system dynamics chaos.”
The study was published in the journal PRX Life. Hertz researchers have explored a robot’s “motivation model” that can mimic the decision-making processes of humans and animals even in the absence of clear reward signals.
The study introduces an artificial intelligence (AI) formula that can calculate a way for a robot to determine future actions without direct instructions or manual input.
Daniel Polani, professor of computer science and senior author, explained: “In an application sense, this could mean letting the robot play and manipulate objects on its own without being told.
“It could enhance the way robots learn to interact with humans and other robots by encouraging more ‘natural’ behaviors and interactions.
“There are further applications, such as the survival behavior of semi-autonomous robots placed in places inaccessible to human operators, such as underground or interstellar locations.”
In humans and animals, one theory postulates that there is an “intrinsic motivation” in which behavior is driven only by interactions between the organism and its environment, rather than by specific learned rewards, such as food. This paper successfully transforms the “intrinsic motivation” theory into a theory that can be used by robot agents.
Professor Polani added: “This work is exciting because we can now implement a mechanism in robots similar to the one that helps humans and animals solve new problems without experience.
“We hope that based on this work, we can develop more humanoid robots with more intuitive processes in the future. This provides a huge opportunity for more complex robots with decision-making processes similar to ours.”
The theory on which this paper is based is called “maximizing empowerment” and has been developed in Hertz for many years. It suggests that by increasing the range of future outcomes, robots will also have better options in the longer future. Importantly, this approach replaces and therefore potentially eliminates traditional reward systems (such as food signals).
Although maximizing empowerment has shown promise, it has not yet been fully understood or widely applied. Most research used to rely on simulations while carefully calculating necessary information about complex systems, while theory remains challenging.
However, this latest innovative study aims to explain why authorization-based motivations can create behaviors similar to organisms, potentially creating robots with more intrinsic motivations; it also provides a significantly improved way to calculate these motivations.
Professor Polani said the next step is to use this breakthrough algorithm to allow robots to learn more about the world and develop new skills for direct learning, recognition and hone in order to play their value in real-world scenarios.
There are links in the video below this video. If you are interested, you can open it and have a look.
Thank you for watching this video. If you like it, please subscribe and like it. thank
Original text:https://techxplore.com/news/2024-09-algorithm-robots-closer-intuition.html
More information: Stas Tiomkin et al., Intrinsic motivations for dynamic control systems, PRX Life (2024). DOI：
Journal information: PRX Life
Oil tubing: