Fortune Views

― Advertisement ―

spot_img

Despite $75B in VC Investments in Q4, Startups Struggle to Secure Funding

Venture capital investment in U.S. startups surged to $74.6 billion in the fourth quarter of 2024, a significant rise from the $42 billion average...
HomeNewsNew AI Training Method Enables Derpy Robots to Showcase Soccer Skills

New AI Training Method Enables Derpy Robots to Showcase Soccer Skills

Humanoid robots have long posed a challenge for scientists due to the intricacies involved in programming every movement. However, advancements in AI are changing the game. Self-learning capabilities powered by AI mean that robots no longer require every single motion to be pre-scripted or programmed. Take, for instance, the Figure 01 robot, which learned to make coffee after just 10 hours of training footage.

In a recent study by Google DeepMind scientists, miniature Robotis OP3 robots were trained to play soccer using a technique called “deep reinforcement learning.” This method combines various AI training techniques, including reinforcement learning, supervised learning, and deep learning using neural networks.

Initially, the robots relied on a scripted baseline of skills before the scientists applied their combined deep reinforcement learning AI training technique. The results were impressive. In matches, the AI-trained robots exhibited significant improvements in speed, agility, ball handling, and recovery from falls compared to robots not trained using this technique.

Specifically, the AI-trained robots walked 181% faster, turned 302% faster, kicked the ball 34% faster, and recovered from falls 63% faster. Moreover, the robots developed “emergent behaviors,” such as pivoting on the corner of their foot and spinning around, which would be extremely challenging to program manually.

These findings demonstrate the effectiveness of AI training techniques in creating simple yet safe motions in humanoid robots. This opens up possibilities for more sophisticated movements in changing and complex situations. As AI continues to evolve, we can expect further advancements in robotics, paving the way for increasingly capable and adaptable humanoid robots.