A misconception is currently thriving in the industry that one can become a Generative AI expert without learning “traditional” machine learning. Large Language Models (LLMs) predict the next word in a sequence of words. They calculate the probability of occurrence of each word in a vocabulary that can follow a sequence of words. In the
M
Sony AI’s Ping Pong Robot Wins Against Elite Players
April 25, 2026 at 4:40 PM • by MLQ Agent
Sony AI has developed an autonomous robot named Ace that defeated elite amateur table tennis players in matches sanctioned by the International Table Tennis Federation. The robot, part of Project Ace, relies on machine learning to compete at a high level in the fast-paced sport[1][3].
Robot Capabilities and Technology
Ace combines high-speed vision systems, advanced AI decision-making, and an eight-jointed robotic arm to perceive, react to, and return shots at elite speeds and spins. It operates without pre-programmed moves, learning through machine learning techniques similar to its predecessor, Gran Turismo Sophy, which excelled in racing simulations[1][3]. The system matches or exceeds human reaction times in real-time physical interactions[3].
Match Results Against Humans
In tests, Ace faced seven players: five elite amateurs averaging 20 hours of weekly practice and two Japanese professional league players, Minami Ando and Kakeru Sone. It won three best-of-three matches against amateurs, securing 7 wins out of 13 games. Against professionals, it won 1 out of 7 games but lost both best-of-five matches[3]. These results surpass previous table tennis robots[3].
Development Background
Sony AI’s Zürich team, led by roboticist Peter Dürr, built Ace over years of research. The project represents progress from virtual AI agents to physical ones capable of expert-level play in dynamic sports. A related study appeared in Nature, detailing the system’s architecture[3][5].
Physical AI Integration Breakthrough
Project Ace demonstrates a leap in physical AI by integrating perception, reasoning, and action in unpredictable real-world settings. Table tennis’s demands—split-second decisions amid speeds up to 100 km/h and variable spins—test limits of current robotics, yet Ace’s 54% win rate against elite amateurs shows viable human-level performance[3]. This success stems from end-to-end learning, where the AI processes raw sensor data directly to motor commands, bypassing rigid programming that plagued earlier robots. The achievement underscores Sony AI’s shift from simulation-based AI, like Gran Turismo Sophy, to embodied agents. Peter Dürr noted it proves robots can win competitive sports by matching human reaction times[3]. Peter Stone, Sony AI chief scientist, emphasized its broader implications for precision tasks beyond sports, highlighting a milestone in scalable AI hardware-software integration[3].
Refinements Against Professionals
Sony AI plans to refine Ace’s techniques against top professionals, targeting consistent wins in extended matches. Future iterations may incorporate adaptive learning during games, enhancing spin prediction and tactical play to close the gap with pros[3]. Expansion to other racket sports or collaborative human-robot scenarios could follow, building on this real-time interaction foundation. Industry-wide, Ace sets a benchmark for fast-action robotics, potentially accelerating applications in manufacturing, healthcare, and sports training. As hardware costs decline, similar systems might enter consumer markets for coaching or entertainment by late 2027. Researchers anticipate open-sourcing elements of the framework to spur innovation in dynamic environments[3].
Written with AI assistance, verified and edited by our team. Questions? Contact us.
Related Posts
![[institutes-now]-pre-training-‘warm-up’-helps-ai-models-recognize-their-own-ignorance](https://sapio.asia/wp-content/uploads/2026/04/37106-institutes-now-pre-training-warm-up-helps-ai-models-recognize-their-own-ignorance.jpg)
