Scientists at the ETH Zurich in Switzerland university have created an AI robot, CyberRunner, that can solve a physical marble labyrinth in record time. This remarkable achievement challenges the notion that the game requires the subtle physical skills and intuition of humans. The AI robot, using deep reinforcement learning, learned to navigate the labyrinth through experience, capturing observations, and receiving rewards based on its game performance. Surprisingly, the machine was built at a cost of under $200, raising hopes for its widespread replication in various fields. During the learning phase, the robot even discovered shortcuts, prompting researchers to intervene and instruct it not to exploit them. Lead researcher Prof. Raffaello D’Andrea expressed excitement about the affordability of the technology, stating that it allows anyone to engage in cutting-edge AI research applied to the physical world. The research paper can be found on the project’s website, CyberRunner.ai.
The AI robot, named CyberRunner, was developed by scientists at the ETH Zurich in Switzerland. Its purpose was to tackle the challenge of solving a physical marble labyrinth, a game that has long been considered the domain of human expertise. This particular game requires subtle physical skills and intuition, setting it apart from purely strategic tasks like chess or Go. However, the researchers managed to teach the AI robot to master the labyrinth using deep reinforcement learning.
What’s even more impressive is that the entire machine was built at a cost of under $200, making it an affordable option for those looking to delve into AI research. The CyberRunner learns through experience, gathering information from its observations and receiving rewards based on its performance in the game. With the help of two knobs that control the orientation of the labyrinth board, the AI robot demonstrates fine motor skills and spatial reasoning.
One of the standout achievements of CyberRunner was its ability to complete the learning process on the real-world labyrinth in just 6.06 hours, surpassing the record time set by skilled human players. Astonishingly, during its learning phase, the robot even discovered shortcuts, showing its advanced problem-solving capabilities. However, researchers intervened and instructed the AI not to exploit these shortcuts, preventing it from cheating.
Lead researcher Prof. Raffaello D’Andrea shared his enthusiasm about the project’s affordability, emphasizing that it allows anyone with an interest in the field to engage in groundbreaking AI research as it applies to the physical world. The research paper detailing the development and capabilities of CyberRunner can be accessed on the project’s website, CyberRunner.ai.
The creation of this AI robot marks a significant milestone in the field of artificial intelligence. The ability to solve a physical marble labyrinth, a task once considered exclusive to humans, highlights the incredible progress made in deep reinforcement learning. By harnessing the power of experience-based learning, the scientists at ETH Zurich have demonstrated that AI is capable of performing intricate tasks that traditionally require human dexterity and intuition.
This breakthrough carries far-reaching implications for various industries. The affordability of the machine opens doors for researchers, enthusiasts, and innovators worldwide to explore novel applications of AI in solving physical challenges. The potential for advancements in robotics, gaming, and even education is immense.
As the research paper becomes widely available, it is likely that other researchers and developers will build upon the work of the ETH Zurich team, further expanding the possibilities and applications of AI in the real world. The CyberRunner project serves as both a testament to human ingenuity and a glimpse into a future where AI and humans collaborate to push the boundaries of what is possible.
