Today, you will find 3D printers in the most surprising places—and all over the world. Not only that, but they are often busy doing the most surprising things for the human race. If you have been following 3D printing for even the shortest amount of time, then you may have learned to continually expect the unexpected. Machine learning and data calculations are perfect examples of this as they are now being applied in 3D via a new artificial intelligence system that performs its work through bending light.
AI is built on looping calculations of numbers and data that ultimately result in recognition. As UCLA researchers discovered, they can actually translate this into physical form through 3D printing with the accompaniment of light. Working with the premise of 3D printing that is built on layers, the researchers are able to make transparent diffraction patterns, ‘reflecting’ data in a neural network. Their work was recently published in a paper titled ‘All-optical machine learning using diffractive deep neural networks,’ by Xing Lin, Yair Rivenson, Nezih T. Yardimci, Muhammed Veli, Yi Luo, Mona Jarrahi, and Aydogan Ozcan.
“Deep learning has been transforming our ability to execute advanced inference tasks using computers,” state the researchers in their abstract.
They are able to establish this with a diffractive deep neural network, also known as a D2NN architecture—and one that is able to perform functions based on the collective diffractive layers.
“We create 3D-printed D2NNs that implement classification of images of handwritten digits and fashion products as well as the function of an imaging lens at terahertz spectrum,” state the researchers. “Our all-optical deep learning framework can perform, at the speed of light, various complex functions that computer-based neural networks can implement, and will find applications in all-optical image analysis, feature detection and object classification, also enabling new camera designs and optical components that perform unique tasks using D2NNs.”
The team used their unique learning model to recognize numbers that were written out by hand, afterward converting the matrix math into a data series related to ‘optical transformations.’ Each layer contributes to creating this data with the use of light, refocusing and adding values. This process involves millions of optical transformations on the 3D printed plates, with the system translating light into numeric values; in fact, the researchers report that with the 3D-printed D2NNs they are having a 90 percent success rate!
While the process is still being refined by the research team, it could be relevant to numerous applications in the future due to the flexibility of such calculation tools. It could be used to read letters instead of numbers too, as well as offering facial or other types of physical recognition. Find out more about this research or order the article for further reading here.
Discuss this article and other 3D printing topics at 3DPrintBoard.com.[Source / Images: TechCrunch]
You May Also Like
3D Printing News Briefs: May 9, 2019
For today’s 3D Printing News Briefs, we’re starting with some business and software news, then moving on to examples of things you can make with 3D printing – even the...
Alchemite Machine Learning Engine Used to Design New Alloy for Direct Laser Deposition 3D Printing
Artificial intelligence (AI) company Intellegens, which is a spin-off from the University of Cambridge, created a unique toolset that can train deep neural networks from noisy or sparse data. The machine...
3D Printing News Briefs: February 22, 2019
We’ve got some exciting dental news to share first in today’s 3D Printing News Briefs – Stratasys just announced its new full-color dental 3D printer at LMT Lab Day. Moving...
Dan Wellers, Digital Futures Lead at SAP “3D Printing will Expand into 4D Printing”
We’re seeing an increased focus by software firms in our teeny tiny industry. Simulation, PLM, CAD, the Digital Twin, Industry 4.0 have some very large and influential firms salivating about that...
View our broad assortment of in house and third party products.