ATEC2025 Challenges Robots in Real-World Extreme Conditions to Bridge Lab-to-Deployment Gap
The ATEC2025 Real-World Extreme Challenge, held at The Chinese University of Hong Kong, tested intelligent robots in unpredictable outdoor environments including mud, stairs, and swaying bridges. Organized with support from Ant Group, the competition aimed to bridge the gap between laboratory demonstrations and actual deployment by evaluating locomotion, manipulation, and environmental modification capabilities. Out of 392 global teams, only 13 finalists competed in tasks such as waste sorting, autonomous plant watering, field orienteering, and bridge crossing. Professor Yunhui Liu, Co-Chair of ATEC2025, emphasized that failure is a normal part of this inaugural format, highlighting progress over polished performances. The event underscores a shift in artificial intelligence towards integrating machine intelligence with the physical world. While the winning team receives a $150,000 prize, the primary goal is foundational advancement rather than immediate commercialization. Liu noted that reliable real-world deployment for complex household tasks remains years away, but achieving high success rates in basic human tasks would be transformative. This competition serves as a critical benchmark for embodied AI, leveraging the Greater Bay Area's ecosystem to foster robust robotic intelligence.
Wire timeline
ATEC2025 Challenges Robots in Real-World Extreme Conditions to Bridge Lab-to-Deployment Gap
The ATEC2025 Real-World Extreme Challenge, held at The Chinese University of Hong Kong, tested intelligent robots in unpredictable outdoor environments including mud, stairs, and swaying bridges. Organized with support from Ant Group, the competition aimed to bridge the gap between laboratory demonstrations and actual deployment by evaluating locomotion, manipulation, and environmental modification capabilities. Out of 392 global teams, only 13 finalists competed in tasks such as waste sorting, autonomous plant watering, field orienteering, and bridge crossing. Professor Yunhui Liu, Co-Chair of ATEC2025, emphasized that failure is a normal part of this inaugural format, highlighting progress over polished performances. The event underscores a shift in artificial intelligence towards integrating machine intelligence with the physical world. While the winning team receives a $150,000 prize, the primary goal is foundational advancement rather than immediate commercialization. Liu noted that reliable real-world deployment for complex household tasks remains years away, but achieving high success rates in basic human tasks would be transformative. This competition serves as a critical benchmark for embodied AI, leveraging the Greater Bay Area's ecosystem to foster robust robotic intelligence.
TechNode