Dive brief:
- A new training method for robots can bring them much closer to performing complex tasks in the workplace with more efficient results, according to a new study written by researchers at the Massachusetts Institute of Technology.
- MIT’s Improbable AI Lab, a group within its Laboratory for Computer Science and Artificial Intelligence, combined three different types of instruction (language-based, visual, and action-based) to teach robots how to perform multi-step tasks with higher success rates than comparable training methods, according to a statement from the institution.
- The training framework, known as Compositional-Based Models for Hierarchical Planning (HiP for short) can help robots not only make an impact in construction, but also perform feats in the manufacturing space, so as in housework, according to the release.
Diving knowledge:
Using the three different types of instruction parameters, robots can approach their “long-term goal,” or an overall goal that involves completing many smaller steps first. These tasks that humans take for granted include making a sandwich, brewing a pot of tea, or moving materials to the right place based on their appearance.
The HiP process works as a hierarchy, with the ability to pre-train each component on different data sets, including information outside of robotics, according to the release.
“Instead of pushing one model to do everything, we combine several that take advantage of different modalities of Internet data,” said Anurag Ajay, Ph.D. student at MIT and a subsidiary of CSAIL, the statement said. “When used in tandem, they aid robotic decision-making and can help with tasks in homes, factories and construction sites.”
Each step is improved by the next through a process called “iterative refinement” that allows HiP to reason about its ideas and receive feedback at each stage to generate a more practical scheme, according to the release. MIT likens this process to an author sending material to their editor, who then returns red ink and notes to the copy.
Robots in the workplace
Ajay told Construction Dive that the robots are currently limited by their hardware, but that they would be good at doing brute force work, such as picking up very heavy objects and moving them from one place to another, and going up or down stairs.
“Humans are pretty good at jobs that require good dexterous manipulation, but some things that have to do with lifting heavy objects and basically moving into places that cranes couldn’t go as an example, those would be . . . .good areas of application,” Ajay. said
this marriage of AI and robotics it’s a trend that contractors are watching closely as we head into 2024. While robots aren’t yet in workplaces everywhere, the technology to make that happen is improving.
HiP isn’t the only type of AI-related training for robots, either: researchers at the Toyota Research Institute are using a generative AI technique known as “diffusion politics” to teach your own robots how to perform complex feats.
Robots aren’t even the full limit of what AI is capable of in the construction industry. In addition to eye-catching apps, it could also help builders save time on email responses, searching for change order information via a chat app, or automating layouts.