Research Bits: April 25

New superconductor category proposed; using uncertainty in AI; 2-stimuli programmable fabric.

popularity

Superconductor breakthrough — palladium
Palladium may be a better superconductor than even nickelates (superconductors based on nickel), according to research by TU Wien working with Japanese universities. The research shows that palladates may be a ‘Goldilocks material’ in which it can continue its superconducting state at a higher temperature.

“Palladium is directly one line below nickel in the periodic table. The properties are similar, but the electrons there are on average somewhat further away from the atomic nucleus and each other, so the electronic interaction is weaker,” said Prof. Karsten Held from the Institute of Solid State Physics at TU Wien in a press release.

According to the researcher’s paper “Optimizing superconductivity: from cuprates via nickelates to palladates,” the theoretically proposed palladates have yet to be synthesized. “The computational results are very promising,” says Held.

Superconductors are used as the microwave frequency filters installed in more than 10,000 mobile communications towers to prevent interference among the thousands of mobile signals, according to the U.S. Department of Energy’s Office of Scientific and Technical Information.

Reference: “Optimizing superconductivity: from cuprates via nickelates to palladates,” Motoharu Kitatani, Liang Si, Paul Worm, Jan M. Tomczak, Ryotaro Arita, Karsten Held. https://doi.org/10.48550/arXiv.2207.14038; PDF available here: https://arxiv.org/pdf/2207.14038v1.pdf.

AI expresses uncertainty — and that’s a good thing
Engineering and health researchers from North Carolina State University are using a cough-detecting AI algorithm to improve cough detection in electronic devices by giving AI the option to be uncertain. Because background noise can often confuse algorithms, the team gave their algorithm the option to say it did not know what sound it was hearing.

“When AI is being trained to identify the sound of coughing, this is usually done with ‘clean’ data – there is not a lot of background noise or confusing sounds,” said Edgar Lobaton, associate professor of electrical and computer engineering at North Carolina State University and author of team’s research paper. Because of all the real-world noise that could confuse algorithms, most cough detection technologies had false reports of what was a cough when trying to pick out the sound out among background noise.

“In the near term, our work limits the reporting of false positives by allowing the AI to note when it hears sounds that it can’t identify,” said Lobaton. “In the longer term, our algorithm should allow us to continually train the AI, by telling it whether the unfamiliar sounds it is hearing are coughs or are unrelated noises. This should allow for much more precise detection over time.” The team also reports that its algorithm uses fewer sound samples per second to ID the cough — 750 versus 16,000 sound samples per second.

Reference: “Robust Cough Detection with Out-of-Distribution Detection,” Yuhan Chen, Pankaj Attri, Jeffrey Barahona, Alper Bozkurt and Edgar Lobaton, North Carolina State University; Michelle L. Hernandez and Delesha Carpenter, University of North Carolina at Chapel Hill. Published: April 5, IEEE Journal of Biomedical and Health Informatics. DOI: 10.1109/JBHI.2023.3264783

Warm, programmable fabric
Researchers at Canada’s University of Waterloo have developed a programmable smart fabric that responds to two different stimuli, heat and electricity. Made out of inexpensively produced polymer nano-composite fiber derived from recycled plastic, the material can change its color and shape. Future uses could be in automotive — such as automotive parts that reshape themselves after a collision. Another use is in winter clothing that can heat the wearer on demand.

Reference: “Multi-Stimuli Dually-Responsive Intelligent Woven Structures with Local Programmability for Biomimetic Applications,” Runxin Xu, Guanzheng Wu, Mengmeng Jiang, Shaojie Cao, Mahyar Panahi-Sarmad, Milad Kamkar, Xueliang Xiao. First published: 19 February 2023. https://doi.org/10.1002/smll.202207900

More research
Check out this weeks’ tech papers.

Catch up on industry research and curated tech papers.

Bookmark our industry-research round-up page.



Leave a Reply


(Note: This name will be displayed publicly)