Researchers have created a Generative Adversarial Network (GAN) for anomaly detection in manufacturing. By deploying a method that clusters data and individually assesses...
Researchers have created a Generative Adversarial Network (GAN) for anomaly detection in manufacturing. By deploying a method that clusters data and individually assesses manifolds for anomalies, this method significantly outperforms other GANs in testing. Within particular datasets, D-AnoGAN can be trained to detect industry-specific anomalies with a very low error rate.
Prof. Mary Kraft and her research group have developed a computational reconstruction strategy to reshape 3D images acquired from depth profiling mode on secondary ion...
Prof. Mary Kraft and her research group have developed a computational reconstruction strategy to reshape 3D images acquired from depth profiling mode on secondary ion mass spectrometry (SIMS). This strategy can enhance understanding of structure function relationship of materials using SIMs (e.g. subcellular biological processes). For samples with nonplanar surfaces, secondary ions detected in the same SIMS depth profiling image, and thus depicted at the same z-position with respect to the surface may be from molecules with different z positions. 3D SIMS image depth correction strategy is needed when both substrate signals and atomic force microscopy data are not available for NanoSIMS depth profiling. The reconstruction strategy accurately captures the basic shape of the cell as well as the surface features, in addition to reducing time for complementary instrumental data collection.
Figure 1. Comparison of 3D 18O-Enrichment 3D SIMS Images of 18O-Cholesterol (left: uncorrected, right: corrected with the reconstruction strategy)
Dr. Gabriel Weaver, Dr. Yardley, and Dr. Emmerich have developed a cyber and physical disruption model for electrical power grid operations that allows stakeholders in the...
Dr. Gabriel Weaver, Dr. Yardley, and Dr. Emmerich have developed a cyber and physical disruption model for electrical power grid operations that allows stakeholders in the industry to anticipate, prepare, and avoid damages from various different disruptive events. This model has been improved with fundamental changes to the cyber layer of the pipeline, allowing for a more diverse cyber layer that is nourished by data collected from DARPA RADICS exercises for cyberattacks against the electrical power grid. This diverse cyber layer is now able to apply information diffusion processes running on real-world communications networks under baseline and disrupted conditions to explicitly represent communications network topologies and associated information diffusion processes across each layer in the network stack. The result of this technology is a brand-new solution for stakeholders in assessing critical infrastructure risks in the electrical power grid sector.
Inventors from the University of Illinois and University of Massachusetts at Amherst have developed a novel way of storing data using chemically modified nucleotides...
Inventors from the University of Illinois and University of Massachusetts at Amherst have developed a novel way of storing data using chemically modified nucleotides combined with DNA base pairs. This system of DNA-based storage creates a novel eleven letter alphabet with room for even more letters that allows for a nearly two-fold increase in the storage density of molecular recorders. Furthermore, the inventors have also developed a novel neural network architecture that sequences these alphabets and constructs with an extremely large accuracy. This entire technology drastically improves upon current DNA-based storage systems and provides them with new directions in molecular storage and computing that have the potential to resolve important cost and implementation issues the industry faces today.
