One end in hand, everything is available. Scan to install Meixi Online.
The Allen Institute for Artificial Intelligence (Ai2) recently launched an innovative artificial intelligence robot system—MolmoAct. This system employs novel technologies to assist robots in achieving precise navigation in complex and chaotic real-world environments. Notably, Ai2 has made all the code, data, and training methods of this system publicly available in an open-source format, promoting transparency and sharing in robotic technology.
The MolmoAct system has the capability to convert two-dimensional images into three-dimensional visual effects, allowing for pre-execution simulations of actual actions and enabling human operators to adjust robot movements in real-time, significantly enhancing the flexibility and accuracy of task execution by robots. Unlike traditional “black box” robot models that rely on proprietary data, MolmoAct emphasizes openness and transparency, paving new avenues for robotic research and applications.
Ai2 hopes that MolmoAct will serve as a solid foundation for robotic researchers, enterprises, and developers, aiding in the creation of robots capable of autonomous operation in unstructured environments, with application scenarios covering complex settings such as homes, warehouses, and disaster rescue. According to GeekWire, during a recent demonstration held at Ai2’s new headquarters on the banks of Lake Washington in Seattle, researchers showcased how robots could be controlled through natural language commands to pick up household items like cups and plush toys and accurately move them to designated locations.
As an important part of Ai2’s open-source AI ecosystem, MolmoAct is based on its multimodal AI model Molmo, adding three-dimensional reasoning and robot motion control capabilities. Ai2 also possesses a fully transparent flagship large language model, OLMo, with publicly available training data, code, and model weights, supporting artificial intelligence research and public oversight.
Ranjay Krishna, head of computer vision at Ai2 and assistant professor at the University of Washington, stated that Ai2’s projects are moving towards a “unified AI model,” aiming to create a comprehensive intelligent agent with capabilities in reasoning, language understanding, visual perception, and robot control.
Exciting Recommendations
US and China Extend Trade Truce for 90 Days 
US Vaccination Rates May Fall Below Immunity Threshold 
Emergency Medical Bills Exceeding $100,000 in 15 Minutes 
Chinese Man Charged with Stealing Rare Manuscripts from UCLA