Lifeboat Foundation

Experts in the AI and Big Data sphere consider October 2021 to be a dark month. Their pessimism isn’t fueled by rapidly shortening days or chilly weather in much of the country—but rather by the grim news from Facebook on the effectiveness of AI in content moderation.

This is unexpected. The social media behemoth has long touted tech tools such as machine learning and Big Data as answers to its moderation woes. As CEO Mark Zuckerberg explained for CBS News, “The long-term promise of AI is that in addition to identifying risks more quickly and accurately than would have already happened, it may also identify risks that nobody would have flagged at all—including terrorists planning attacks using private channels, people bullying someone too afraid to report it themselves, and other issues both local and global.”

None.

Scientists from Heidelberg and Bern have succeeded in training spiking neural networks to solve complex tasks with extreme energy efficiency. The advance was enabled by the BrainScaleS-2 neuromorphic platform, which can be accessed online as part of the EBRAINS research infrastructure.

Developing a machine that processes information as efficiently as the human brain has been a long-standing research goal towards true artificial intelligence. An interdisciplinary research team at Heidelberg University and the University of Bern led by Dr Mihai Petrovici is tackling this problem with the help of biologically-inspired artificial neural networks.

Continue reading “Human Brain Project researchers demonstrate highly efficient deep learning on a spiking neuromorphic chip” »

This is the first installment blog summarizing AI Theology’s panel discussion on how to make Facebook algorithms accountable to users.

A commonly available oral diuretic pill approved by the U.S. Food and Drug Administration may be a potential candidate for an Alzheimer’s disease treatment for those who are at genetic risk, according to findings published in Nature Aging. The research included analysis showing that those who took bumetanide — a commonly used and potent diuretic — had a significantly lower prevalence of Alzheimer’s disease compared to those not taking the drug. The study, funded by the National Institute on Aging (NIA), part of the National Institutes of Health, advances a precision medicine approach for individuals at greater risk of the disease because of their genetic makeup.

The research team analyzed information in databases of brain tissue samples and FDA-approved drugs, performed mouse and human cell experiments, and explored human population studies to identify bumetanide as a leading drug candidate that may potentially be repurposed to treat Alzheimer’s.

“Though further tests and clinical trials are needed, this research underscores the value of big data-driven tactics combined with more traditional scientific approaches to identify existing FDA-approved drugs as candidates for drug repurposing to treat Alzheimer’s disease,” said NIA Director Richard J. Hodes, M.D.

In Optica, The Optical Society’s (OSA) journal for high impact research, Qiu and colleagues describe a new approach for digitizing color. It can be applied to cameras and displays — including ones used for computers, televisions and mobile devices — and used to fine-tune the color of LED lighting.

“Our new approach can improve today’s commercially available displays or enhance the sense of reality for new technologies such as near-eye-displays for virtual reality and augmented reality glasses,” said Jiyong Wang, a member of the PAINT research team. “It can also be used to produce LED lighting for hospitals, tunnels, submarines and airplanes that precisely mimics natural sunlight. This can help regulate circadian rhythm in people who are lacking sun exposure, for example.”

The company believes its Codex machine learning algorithm is the next step in programming—a sidekick for coders to speed up the work and ease the drudgery.

Topology in optics and photonics has been a hot topic since 1,890 where singularities in electromagnetic fields have been considered. The recent award of the Nobel prize for topology developments in condensed matter physics has led to renewed surge in topology in optics with most recent developments in implementing condensed matter particle-like topological structures in photonics. Recently, topological photonics, especially the topological electromagnetic pulses, hold promise for nontrivial wave-matter interactions and provide additional degrees of freedom for information and energy transfer. However, to date the topology of ultrafast transient electromagnetic pulses had been largely unexplored.

In their paper published in the journal Nature Communications, physicists in the UK and Singapore report a new family of electromagnetic pulses, the exact solutions of Maxwell’s equation with toroidal topology, in which topological complexity can be continuously controlled, namely supertoroidal topology. The electromagnetic fields in such supertoroidal pulses have skyrmionic structures as they propagate in free space with the speed of light.

Skyrmions, sophisticated topological particles originally proposed as a unified model of the nucleon by Tony Skyrme in 1,962 behave like nanoscale magnetic vortices with spectacular textures. They have been widely studied in many condensed matter systems, including chiral magnets and liquid crystals, as nontrivial excitations showing great importance for information storing and transferring. If skyrmions can fly, open up infinite possibilities for the next generation of informatics revolution.

Track code: TD-3

Abstract:
Solar Sails are at the same stage of engineering development as electric motors were in the 1830’s. Each attribute of solar flux has been examined in isolation, such as photon, proton, plasma, and electrodynamic systems. This talk recommends designing a simple baseline system that converges multiple propulsion methods into optimized systems, as is currently done with electric motors. Many convergences can come from this solution space. Once a baseline design is created, AI genetic algorithms can “flight test” and refine the designs in simulation to adjust proportions and geometry. Once a base design is refined, a second AI evolution pass would design fleet systems that flock like birds to optimize performance. These could fly as a protective shield around Mars crewed fleets, provide space based solar power, deploy rapid reaction probes for interstellar comets, and be used in NEO asteroid mining. In the long term, fleets of solar energy management vehicles can provide orbital Carrigan event protection and Martian solar wind protection for terraforming. This talk is also a case study in how technology revolutions happen, and how to accelerate the creation and democratization of technical solutions.

Continue reading “FlexSail: Solar Sails and Tech Revolutions — Kent Nebergall — 2021 Mars Society Virtual Convention” »

Circa 2019 😀

Because they can process massive amounts of data, computers can perform analytical tasks that are beyond human capability. Google, for instance, is using its computing power to develop AI algorithms that construct two-dimensional CT images of lungs into a three-dimensional lung and look at the entire structure to determine whether cancer is present. Radiologists, in contrast, have to look at these images individually and attempt to reconstruct them in their heads. Another Google algorithm can do something radiologists cannot do at all: determine patients’ risk of cardiovascular disease by looking at a scan of their retinas, picking up on subtle changes related to blood pressure, cholesterol, smoking history and aging. “There’s potential signal there beyond what was known before,” says Google product manager Daniel Tse.

The Black Box Problem

Continue reading “Rise of Robot Radiologists” »