last posts

The missing piece of artificial intelligence engineering

techsm5

DfAI: the missing piece of artificial intelligence engineering

DfAI Principles. Credit: Carnegie Mellon University, College of Engineering

Considering how quickly engineering design and manufacturing have advanced alongside computer developments, you may be surprised that very few engineers are educated in both engineering systems design and intelligence. artificial. There are countless opportunities for game-changing improvements in the way we develop new technologies using AI in engineering design, but to succeed in these challenging areas, engineers must understand a new specialty: designing for intelligence. artificial.

Chris McComb, an associate professor of mechanical engineering at Carnegie Mellon, and his student Glen Williams, now a principal investigator at Re:Build Manufacturing, developed a design framework for artificial intelligence (DfAI) in collaboration with researchers from the Penn State University to educate and encourage the academic and industrial engineering community to embrace AI engineering design.

“Most of the time, we think of AI as a tool to add to an existing system, but to develop better systems, we need to integrate AI into the engineering design process from the start,” says McComb.

One of the main challenges is motivating institutions to invest in the long-term potential of AI technologies. Since engineering is product-driven and design and manufacturing incentives prioritize short-term excellence, budgeting resources for long-term research and development is difficult but worthwhile. .

To illustrate the importance of this, Williams described two hypothetical companies mass-producing electric aircraft. For initial development, Company A chooses a manual manufacturing route to get to market quickly and achieve profitability. On the other hand, Company B is building a data-rich process that captures intelligence throughout the design lifecycle. Over the next ten years, Company B is able to significantly reduce its operating costs by using data-driven design that can both optimize its aircraft production and create better products. Company A can no longer follow.

Because design and manufacturing don’t happen in silos, DfAI applies to the broader aspects of the engineering design process. Williams suggests that, fundamentally, advancing DfAI can be approached by 1) increasing AI literacy in industry; 2) redesign engineering systems to better integrate with AI; and 3) improve the process of developing AI in engineering.

“Engineering data is complex and not always tied to the wider community,” says McComb, as one of the reasons other fields can innovate AI technology faster. “The number of experts able to interpret this data is small, so the DfAI will require individuals to have specific expertise. Academia and industry need to work together to support long-term innovation in this area.”

The team describes three people as necessities for DfAI: engineering designers, design repository custodians, and AI developers. An engineering designer can be a person or team responsible for developing specifications for a new project. They are problem solvers who can understand engineering constraints as well as AI algorithms. Design repository custodians need to go deeper into the database maintainer role by having engineering design and manufacturing knowledge to provide design engineers with the data management tools to meet workflow demands and be expandable to future demands. Finally, AI developers must be able to imagine, develop, market and continuously improve AI software products to help design engineers.

“We can’t view AI development as an afterthought in our core operations,” McComb sums up. “Unless we empower design engineers with next-generation design and manufacturing software based on deeply embedded artificial intelligence, our ability to design new and useful technology will fall short of the potential of these new manufacturing techniques. .

According to Williams, several industries may have an easier time adopting the DfAI principles. IT applications, such as products made by digital techniques such as additive manufacturing, naturally have a complex cyber-physical pathway and produce personnel who would be well suited to adopt and apply DfAI principles. Regulated industries such as aerospace and medical devices are accustomed to following rigorous procedures and reliable data storage techniques in the context of highly complex engineering systems, so they likely have the resources to begin the process. adoption of the DfAI.

We could also see the Internet of Things (IoT) and smart device product designers adopt DfAI principles sooner rather than later. These designers will not only benefit from the data during the design process, but also from the large amounts of data collected from their devices during testing or practice. Harnessing this valuable product data in the field could bring huge benefits to AI tools that help improve the quality, performance, durability and profitability of future products.

“Since there is a wide variety of engineering applications, industries, technologies, and scales of operation, establishing general frameworks, common terminology, and written principles is essential to developing a interconnected community of AI engineers who can collaborate together,” says Williams. “Our DfAI framework provides the high-level starting point for these critical discussions.”

More information:
Glen Williams et al, Design for Artificial Intelligence: Proposing a Conceptual Framework Grounded in Data Wrangling, Journal of Computer and Information Science in Engineering (2022). DOI: 10.1115/1.4055854

Provided by Carnegie Mellon University Mechanical Engineering

Quote: DfAI: The missing piece of artificial intelligence engineering (2023, January 3) retrieved January 3, 2023 from https://techxplore.com/news/2023-01-dfai-piece-artificial-intelligence.html

This document is subject to copyright. Except for fair use for purposes of private study or research, no part may be reproduced without written permission. The content is provided for information only.


techsm5

Comments



Font Size
+
16
-
lines height
+
2
-