Lifeboat Foundation

Artificial intelligence (AI) experts at the University of Massachusetts Amherst and the Baylor College of Medicine report that they have successfully addressed what they call a “major, long-standing obstacle to increasing AI capabilities” by drawing inspiration from a human brain memory mechanism known as “replay.”

First author and postdoctoral researcher Gido van de Ven and principal investigator Andreas Tolias at Baylor, with Hava Siegelmann at UMass Amherst, write in Nature Communications that they have developed a new method to protect—” surprisingly efficiently”— deep neural networks from “catastrophic forgetting;” upon learning new lessons, the networks forget what they had learned before.

Siegelmann and colleagues point out that deep neural networks are the main drivers behind recent AI advances, but progress is held back by this forgetting.

Artificial intelligence researchers at North Carolina State University have improved the performance of deep neural networks by combining feature normalization and feature attention modules into a single module that they call attentive normalization (AN). The hybrid module improves the accuracy of the system significantly, while using negligible extra computational power.

“Feature normalization is a crucial element of training deep neural networks, and feature attention is equally important for helping networks highlight which features learned from raw data are most important for accomplishing a given task,” says Tianfu Wu, corresponding author of a paper on the work and an assistant professor of electrical and computer engineering at NC State. “But they have mostly been treated separately. We found that combining them made them more efficient and effective.”

To test their AN module, the researchers plugged it into four of the most widely used neural network architectures: ResNets, DenseNets, MobileNetsV2 and AOGNets. They then tested the networks against two industry standard benchmarks: the ImageNet-1000 classification benchmark and the MS-COCO 2017 object detection and instance segmentation benchmark.

There are a variety of complementary observations that could be used in the search for life in extraterrestrial settings. At the molecular scale, patterns in the distribution of organics could provide powerful evidence of a biotic component. In order to observe these molecular biosignatures during spaceflight missions, it is necessary to perform separation science in situ. Microchip electrophoresis (ME) is ideally suited for this task. Although this technique is readily miniaturized and numerous instruments have been developed over the last 3 decades, to date, all lack the automation capabilities needed for future missions of exploration. We have developed a portable, automated, battery-powered, and remotely operated ME instrument coupled to laser-induced fluorescence detection. This system contains all the necessary hardware and software interfaces for end-to-end functionality. Here, we report the first application of the system for amino acid analysis coupled to an extraction unit in order to demonstrate automated sample-to-data operation. The system was remotely operated aboard a rover during a simulated Mars mission in the Atacama Desert, Chile. This is the first demonstration of a fully automated ME analysis of soil samples relevant to planetary exploration. This validation is a critical milestone in the advancement of this technology for future implementation on a spaceflight mission.

We tackle the crucial challenge of fusing different modalities of features for multimodal sentiment analysis. Mainly based on neural networks, existing approaches largely model multimodal interactions in an implicit and hard-to-understand manner. We address this limitation with inspirations from quantum theory, which contains principled methods for modeling complicated interactions and correlations. In our quantum-inspired framework, the word interaction within a single modality and the interaction across modalities are formulated with superposition and entanglement respectively at different stages. The complex-valued neural network implementation of the framework achieves comparable results to state-of-the-art systems on two benchmarking video sentiment analysis datasets. In the meantime, we produce the unimodal and bimodal sentiment directly from the model to interpret the entangled decision.

Figures 4 and 5 effectively demonstrate that local acceleration is capable of heating electrons to ~7 MeV as the phase space density profiles show signatures of local acceleration during both of the geomagnetic storms considered. The phase space density enhancements for higher energies followed the enhancements at lower energies. In Supplementary Note 8, additional analysis establishes that locally growing peaks are also observed for lower values of K, corresponding to radiation belt electrons confined closer to the equator. Furthermore, as the values of K and L* are dependent on the magnetic field model chosen, results using an additional two field models are also presented (see Supplementary Note 9) and, once again, growing peaks are observed in the radial phase space density profile. Our results demonstrate that local acceleration had a significant effect on radiation belt particles during both of the storms in October 2012, acting on electrons up to 7 MeV. In the radiation belt region, local acceleration introduces radial gradients in phase space density and so is always accompanied by both outwards and inwards radial diffusion. Locally heating electrons to ~7 MeV provides a very high energy “source population” for inwards radial diffusion and could therefore help explain the occurrence of ~10 MeV electrons in April–May 2017¹⁶.

A recent study by Zhao et al.¹⁵, considered the acceleration of ultra-relativistic electrons via a statistical analysis of events during the Van Allen Probe era. The results were consistent with a two-step acceleration process, where locally heated electrons at large L*, beyond the Van Allen Probes apogee, are radially diffused inwards to reach energies of 7 MeV in the outer radiation belt. While the combination of local acceleration and radial diffusion produces 7 MeV enhancements¹⁵, the Van Allen Probe observations for the two storms shown in this study demonstrate that local acceleration can also act directly up to 7 MeV energies. The local energization mechanism responsible for generating 7 MeV electrons in the heart of the outer radiation belt, be that acceleration by chorus waves or some other process, presents an interesting focus for future research. Longer term analysis and statistical studies can be used to better understand the conditions leading to acceleration. Datasets formed via data-assimilation techniques may be useful for this purpose. Long term observations of the ultra-relativistic component of Earth’s radiation belts demonstrate that ≥7 MeV electrons are a relatively rare phenomenon, occurring far less frequently than enhancements at 1 or 2 MeV¹. It therefore follows that the circumstances leading to multi-MeV enhancements could be unusual, requiring specific conditions. Our results highlight that wave-particle interactions can provide the primary acceleration mechanism for electrons up to ultra-relativistic energies, a finding applicable to magnetized plasmas throughout the solar system.

Chlorine isotopes are a sensitive tracer of degassing throughout planetary evolution that provide evidence for the universal depletion of volatiles in the Moon. We show that much of the chlorine in mare basalts is trapped in water-soluble phases from vapor deposition with low isotope values, with the remaining being isotopically heavy from degassing. We also use halogen concentrations and bulk-Cl isotope values to show that most lunar halogen loss and heavy Cl enrichment occurred during the Giant Impact—resulting in a 10× depletion of halogens relative to the Earth. Last, we conclude that lunar apatite has much higher δ³⁷ Cl values compared to the bulk rock, likely explained by localized degassing, making their use as direct probes of planetary-scale processes problematic.

Lunar mare basalts are depleted in F and Cl by approximately an order of magnitude relative to mid-ocean ridge basalts and contain two Cl-bearing components with elevated isotopic compositions relative to the bulk-Earth value of ∼0‰. The first is a water-soluble chloride constituting 65 ± 10% of total Cl with δ³⁷ Cl values averaging 3.0 ± 4.3‰. The second is structurally bound chloride with δ³⁷ Cl values averaging 7.3 ± 3.5‰. These high and distinctly different isotopic values are inconsistent with equilibrium fractionation processes and instead suggest early and extensive degassing of an isotopically light vapor. No relationship is observed between F/Cl ratios and δ³⁷ Cl values, which suggests that lunar halogen depletion largely resulted from the Moon-forming Giant Impact. The δ³⁷ Cl values of apatite are generally higher than the structurally bound Cl, and ubiquitously higher than the calculated bulk δ³⁷ Cl values of 4.1 ± 4.0‰.

Techniques for imaging through scattering media are generally invasive, operate at microscopic scales or require a priori information. Here, the authors overcome these limitations by introducing confocal diffuse tomography, which captures the 3D shape of objects hidden behind scattering media.

Structural coloration techniques have improved display science due to their high durability in terms of resistance to bleaching and abrasion, and low energy consumption. Here, we propose and demonstrate an all-solid-state, large-area, lithography-free color filter that can switch structural color based on a doped semiconductor. Particularly, an indium-gallium-zinc-oxide (IGZO) thin film is used as a passive index-changing layer. The refractive index of the IGZO layer is tuned by controlling the charge carrier concentration; a hydrogen plasma treatment is used to control the conductivity of the IGZO layer. In this paper, we verify the color modulation using finite difference time domain simulations and experiments. The IGZO-based color filter technology proposed in this study will pave the way for charge-controlled tunable color filters displaying a wide gamut of colors on demand.

© 2020 Chinese Laser Press

Do you know that 1 Starship can carry 100 passengers at a time to MARS!!
But how many would be needed for million people??
Watch yourself!!
#ElonMusk
#SpaceX
#MarsExploration
#SpaceExploration

Do you know that 1 Starship can carry 100 passengers at a time to MARS!!

But how many would be needed for million people?? Watch yourself!! #ElonMusk #SpaceX #MarsExploration #SpaceExploration