Tomato-Picking Robot Conceptualized by Researchers in Collaboration with Chat-GPT

EPFL Researchers Use Chat-GPT to Design First Working Robotic Tomato-Harvester

The development of large language models or LLMs has attracted significant attention due to the ability of these models to process large volumes of text data and answer prompts. These neural networks, such as Chat-GPT, provide immense potential for transforming the way we produce, learn, and even create art. The latest application of the technology by EPFL researchers is to the field of robotic design.

In a case study published in Nature Machine Intelligence, the head of EPFL’s Computational Robot Design & Fabrication Lab, Josie Hughes, together with PhD student Francesco Stella and Cosimo Della Santina of TU Delft, used Chat-GPT to design a fully functional robotic tomato-harvester. The study provides a framework for collaborative design of such devices by humans and LLMs.

According to Hughes, “Even though Chat-GPT is a language model and its code generation is text-based, it provided significant insights and intuition for physical design and showed great potential as a sounding board to stimulate human creativity”. The researchers describe the opportunities and risks involved in using AI tools in robotics and argue that it could transform the design of robots, thereby simplifying and enhancing the process.

The study involved two phases: the first phase was an ideation discussion between the researchers and LLM to define the robot’s purpose, design parameters, and specifications. The second phase involved realizing the robot in the real world, which included refining the LLM-generated code, fabricating the device, and debugging its functionality.

During the first phase, the researchers engaged the LLM in high-level conceptual conversations, exploring the potential challenges facing humanity and using the LLM’s access to global data to offer the “most probable” answer to certain prompts, such as “what features should a robot harvester have?” Once the robot’s basic format was established, the researchers could then pose more specific questions on technical design, materials, and computer code for controlling the device.

The researchers have outlined several potential human-LLM collaboration modes, such as collaborative exploration, which uses AI to augment researchers’ expertise by contributing wide-ranging knowledge beyond their fields. AI can also act as a “funnel,” refining the design process and providing technical input, with humans retaining creative control.

Despite the vast potential of LLMs, there are ethical and logical risks associated with their use, such as bias, plagiarism, and intellectual property issues, as it’s uncertain whether an LLM-generated design can be considered novel. Hughes cautioned, “When decisions are made outside the scope of knowledge of the engineer, this can lead to significant ethical, engineering, or factual errors.”

However, based on their experience, Hughes and her team conclude that LLMs can be a force for good if well managed. “The robotics community must, therefore, identify how to leverage these powerful tools to accelerate the advancement of robots in an ethical, sustainable and socially empowering way,” the researchers say.

Editor Notes:

The use of Chat-GPT in robotic design provides a glimpse into the future potential of AI in the engineering industry. It is crucial to recognize that the role of AI must be carefully evaluated to mitigate ethical and logical risks. However, EPFL’s groundbreaking study suggests that LLMs like Chat-GPT could revolutionize the field of robotics design and simplify the process while enhancing creativity. For more updates and news on the latest AI applications, check out GPT News Room.

Source link

from GPT News Room https://ift.tt/xLz2lSI