Share This Article
The Robotics Transformer 2 AI model is capable of controlling actions
Google Corporation presented its new development – Robotics Transformer 2. This is an artificial intelligence model with a visual language trained with information from the Internet. The robot can perform various actions based on the learned text and images.
The uniqueness of RT-2
The new AI model will enable robots to perform simple tasks like picking up rubbish. Such actions seem simple to humans, but computers need to develop an algorithm that consists of many actions. The robot is able to build them on its own, based on the information it learns from online training. Robotics AI eliminates the need to program the machine to perform each action. The robot can build them on its own based on information it learns from online training.
Robotics Transformer 2 is a unique AI model of the VLA type. It incorporates visual and linguistic components and actions through which it directly controls the robot. RT-2 works by processing information from the web and sending it to the robot to guide it.
According to Google, the new generation has several advantages over the previous one:
- the performance of the RT-2 is almost double that of the first generation;
- the improved model can respond to user commands and translate its own reasoning, albeit at an elementary level.
With the ability to make inferences, RT-2 can decide which action is optimal for a given situation, such as which item to use when cleaning.
Other AI developments
IT giant Google is one of the leading developers in the field of AI. In 2023, it unveiled RoboCat, an AI robotic solution that knows how to improve itself. This AI agent is capable of analysing and remembering different tasks used in different contexts. This gives it new data to learn from and improve its algorithms. RoboCat can significantly improve its knowledge of robotics in just 100 information demonstrations. At the same time, it does not require strict human supervision.
Google is thus getting closer to creating adaptable robots that can perform various actions and learn independently. The company is trying to integrate AI for use in robotics. It is also focusing on teaching artificial intelligence to process information in a similar way to how humans think. This will help make robots as useful to society as possible.
In general, we can talk about the high potential of VLA models for practical applications, but it is too early to talk about the development of human-centred robots.