"learning performance"

#13457708

16 Mar 2026

The MagicBot Z1, a high-performance humanoid robot developed by MagicLab, is pictured in a resting state suspended from a specialized support rack at the company's exhibition stand during the Mobile World Congress in Barcelona, Spain, on March 5, 2026. This advanced bipedal platform is powered by the NVIDIA Jetson AGX Orin 64GB edge computing module, which acts as its primary brain for real-time motion control and spatial perception. The robot's sensory head unit is equipped with a multimodal vision system featuring a 12-megapixel wide-angle RGB camera for object recognition and dual iToF (indirect Time-of-Flight) depth sensors for precision navigation. Its mechanical structure utilizes high-torque quasi-direct drive (QDD) actuators, allowing for a payload capacity of 5 kg per arm and a maximum walking speed of 5 km/h. The display highlights the Z1's integrated end-to-end neural network architecture, which enables it to learn complex tasks such as item sorting and tool handling through human-in-the-loop imitation learning.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13457274

16 Mar 2026

A humanoid robot, identified as the AgiBot X2, performs a synchronized dance routine at the China Mobile pavilion during the Mobile World Congress in Barcelona, Spain, on March 5, 2026. Developed in collaboration with China Mobile, this advanced service robot stands 1.27 meters tall and is powered by a high-performance NVIDIA Jetson Thor computing platform designed for embodied AI. The hardware features over 30 active degrees of freedom and high-torque density actuators, allowing for complex movements such as hip-hop dancing, Tai Chi, and full splits. Its sensor suite includes a 3D LiDAR for spatial mapping, stereo depth cameras for real-time obstacle avoidance, and a multi-microphone array for localized voice interaction. The demonstration highlights China Mobile's 5G-Advanced network capabilities, utilizing ultra-low latency and high-bandwidth connectivity to manage the robot's Vision-Language-Action AI model, enabling interaction with visitors and adaptation of its choreography to ambient music in real-time.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13457275

16 Mar 2026

A humanoid robot, identified as the AgiBot X2, performs a synchronized dance routine at the China Mobile pavilion during the Mobile World Congress in Barcelona, Spain, on March 5, 2026. Developed in collaboration with China Mobile, this advanced service robot stands 1.27 meters tall and is powered by a high-performance NVIDIA Jetson Thor computing platform designed for embodied AI. The hardware features over 30 active degrees of freedom and high-torque density actuators, allowing for complex movements such as hip-hop dancing, Tai Chi, and full splits. Its sensor suite includes a 3D LiDAR for spatial mapping, stereo depth cameras for real-time obstacle avoidance, and a multi-microphone array for localized voice interaction. The demonstration highlights China Mobile's 5G-Advanced network capabilities, utilizing ultra-low latency and high-bandwidth connectivity to manage the robot's Vision-Language-Action AI model, enabling interaction with visitors and adaptation of its choreography to ambient music in real-time.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13457276

16 Mar 2026

A humanoid robot, identified as the AgiBot X2, performs a synchronized dance routine at the China Mobile pavilion during the Mobile World Congress in Barcelona, Spain, on March 5, 2026. Developed in collaboration with China Mobile, this advanced service robot stands 1.27 meters tall and is powered by a high-performance NVIDIA Jetson Thor computing platform designed for embodied AI. The hardware features over 30 active degrees of freedom and high-torque density actuators, allowing for complex movements such as hip-hop dancing, Tai Chi, and full splits. Its sensor suite includes a 3D LiDAR for spatial mapping, stereo depth cameras for real-time obstacle avoidance, and a multi-microphone array for localized voice interaction. The demonstration highlights China Mobile's 5G-Advanced network capabilities, utilizing ultra-low latency and high-bandwidth connectivity to manage the robot's Vision-Language-Action AI model, enabling interaction with visitors and adaptation of its choreography to ambient music in real-time.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13457277

16 Mar 2026

A humanoid robot, identified as the AgiBot X2, performs a synchronized dance routine at the China Mobile pavilion during the Mobile World Congress in Barcelona, Spain, on March 5, 2026. Developed in collaboration with China Mobile, this advanced service robot stands 1.27 meters tall and is powered by a high-performance NVIDIA Jetson Thor computing platform designed for embodied AI. The hardware features over 30 active degrees of freedom and high-torque density actuators, allowing for complex movements such as hip-hop dancing, Tai Chi, and full splits. Its sensor suite includes a 3D LiDAR for spatial mapping, stereo depth cameras for real-time obstacle avoidance, and a multi-microphone array for localized voice interaction. The demonstration highlights China Mobile's 5G-Advanced network capabilities, utilizing ultra-low latency and high-bandwidth connectivity to manage the robot's Vision-Language-Action AI model, enabling interaction with visitors and adaptation of its choreography to ambient music in real-time.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13457278

16 Mar 2026

A humanoid robot, identified as the AgiBot X2, performs a synchronized dance routine at the China Mobile pavilion during the Mobile World Congress in Barcelona, Spain, on March 5, 2026. Developed in collaboration with China Mobile, this advanced service robot stands 1.27 meters tall and is powered by a high-performance NVIDIA Jetson Thor computing platform designed for embodied AI. The hardware features over 30 active degrees of freedom and high-torque density actuators, allowing for complex movements such as hip-hop dancing, Tai Chi, and full splits. Its sensor suite includes a 3D LiDAR for spatial mapping, stereo depth cameras for real-time obstacle avoidance, and a multi-microphone array for localized voice interaction. The demonstration highlights China Mobile's 5G-Advanced network capabilities, utilizing ultra-low latency and high-bandwidth connectivity to manage the robot's Vision-Language-Action AI model, enabling interaction with visitors and adaptation of its choreography to ambient music in real-time.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13457709

16 Mar 2026

The MagicBot Z1 (Quadruped Edition), an advanced robotic dog developed by Magic Lab, is seen in a powered-down resting state suspended from a technical support gantry at the company's innovation booth during the Mobile World Congress in Barcelona, Spain, on March 5, 2026. The robot is built upon a high-performance NVIDIA Jetson AGX Orin 64GB computing module, which processes complex locomotion algorithms and real-time environmental mapping. Its sensor suite, located in the front ''head'' section, features a 12-megapixel ultra-wide RGB camera for visual recognition and a Solid-State LiDAR unit with a 120-degree field of view for obstacle avoidance. The Z1's limbs are powered by proprietary Magic Lab high-torque actuators, providing a peak torque of 180 Nm to each joint, allowing for dynamic movements such as backflips and stair climbing. The demonstration emphasizes the integration of Generative AI for natural language command processing and its IP67-rated chassis, designed for inspection tasks in hazardous industrial environments.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13457220

16 Mar 2026

The Cerebras Systems logo is displayed on a computer screen photographed using a kaleidoscopic filter in Creteil, France, on March 16, 2026. The startup is finalizing its partnership with Amazon Web Services to accelerate the inference of AI models in the cloud.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13403031

4 Mar 2026

The logo of an AI data center, a facility optimized for high-performance computing, deep learning, and training large language models (LLMs), features powerful infrastructure, including advanced liquid cooling, NVIDIA H100 clusters, and high-bandwidth networking. These elements support the immense computational demands of Artificial Intelligence development and deployment. This is exhibited, highlighting the nexus of processing power and neural networks, during the Mobile World Congress (MWC) in Barcelona, Spain, on March 4, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13403034

4 Mar 2026

The specialized AI data center server racks integrate high-density infrastructure optimized for massive scale computing and prominently display distinct logos for their core subsystems: the Liquid Cooling System for efficient thermal management, an AI Processor optimized for deep learning, and AIDC Infra Manager for centralized operational control and monitoring of the advanced computing environment, during the Mobile World Congress (MWC) in Barcelona, Spain, on March 4, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13403071

4 Mar 2026

An expert developer engages in a real-time vocal dialogue with a photorealistic AI avatar, a ''robot girl'' rendered with sub-millisecond latency, against a background of high-density server racks. The scene highlights the seamless integration between LLM Training clusters and Inference Servers, demonstrating the closed-loop feedback where human interaction data is used to fine-tune the model's emotional intelligence and conversational nuance during the Mobile World Congress (MWC) in Barcelona, Spain, on March 4, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13378064

27 Feb 2026

The Baidu logo is displayed on a mobile phone screen in this photo illustration in Brussels, Belgium, on February 27, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13378065

27 Feb 2026

The Baidu logo is displayed on a mobile phone screen in this photo illustration in Brussels, Belgium, on February 27, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13378066

27 Feb 2026

The Baidu logo is displayed on a mobile phone screen in this photo illustration in Brussels, Belgium, on February 27, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13378067

27 Feb 2026

The Baidu logo is displayed on a mobile phone screen in this photo illustration in Brussels, Belgium, on February 27, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.

#13378068

27 Feb 2026

The Baidu logo is displayed on a mobile phone screen in this photo illustration in Brussels, Belgium, on February 27, 2026.

Restricted to Editorial Use Only.
Commercial use is not permitted without prior authorization.
Please contact us for more information.