Lifeboat Foundation

Back in 1993, AI pioneer Jürgen Schmidhuber published the paperA Self-Referential Weight Matrix, which he described as a “thought experiment… intended to make a step towards self-referential machine learning by showing the theoretical possibility of self-referential neural networks whose weight matrices (WMs) can learn to implement and improve their own weight change algorithm.” A lack of subsequent practical studies in this area had however left this potentially impactful meta-learning ability unrealized — until now.

In the new paper A Modern Self-Referential Weight Matrix That Learns to Modify Itself, a research team from The Swiss AI Lab, IDSIA, University of Lugano (USI) & SUPSI, and King Abdullah University of Science and Technology (KAUST) presents a scalable self-referential WM (SRWM) that leverages outer products and the delta update rule to update and improve itself, achieving both practical applicability and impressive performance in game environments.

The proposed model is built upon fast weight programmers (FWPs), a scalable and effective method dating back to the ‘90s that can learn to memorize past data and compute fast weight changes via programming instructions that are additive outer products of self-invented activation patterns, aka keys and values for self-attention. In light of their connection to linear variants of today’s popular transformer architectures, FWPs are now witnessing a revival. Recent studies have advanced conventional FWPs with improved elementary programming instructions or update rules invoked by their slow neural net to reprogram the fast neural net, an approach that has been dubbed the “delta update rule.”

Using material modelling and 3D printing, researchers were able to produce new material quicker and at a lower cost.

10 people will take the better part of a year to port a new technology library. Now we can do it with a couple of GPUs running for a few days.

Nvidia has been quick to hop on the artificial intelligence bus一with many of its consumer facing technologies, such as Deep Learning Super Sampling (DLSS) and AI-accelerated denoising exemplifying that. However, it has also found many uses for AI in its silicon development process and, as Nvidia’s chief scientist Bill Dally said in a GTC conference, even designing new hardware.

Dally outlines a few use cases for AI in its own development process of the latest and greatest graphic cards (among other things), as noted by HPC Wire.

Continue reading “Nvidia’s GPU-powered AI is creating chips with ‘better than human design’” »

Researchers have identified a key regulator of the immune system and are looking to reveal how this protein is regulated to understand how cancer patients have different responses to immunotherapy.

The iconic Hubble Space Telescope observed five galaxies locked in a gravitational dance, before eventually merging together, just before its 23nd anniversary.

Obesity and its consequences on overall health have become a serious global problem. However, very few substances or drugs can fight obesity. In a recent study, Japanese researchers investigated the anti-obesity effects of tocotrienols, a subtype of vitamin E. Experiments showed that tocotrienols significantly reduce the weight gained by mice on a high-fat diet, while also lowering “bad” cholesterol levels. The results pave the way to efficient treatments for obesity.

Over the past few decades, obesity has become increasingly common throughout the entire world. Since obesity often causes other diseases, including diabetes and various cardiovascular disorders, it represents an alarming social problem in both developed and developing countries. Although most of us know that a balanced diet and plenty of exercise are the best ways to prevent obesity, having drugs to effectively treat it would still come in handy if necessary.

Unfortunately, there are very few known substances or drugs that can help prevent or treat obesity and its dreaded secondary diseases. One problem is that some of the mechanisms of obesity-induced diseases are unclear. For example, there is some evidence showing that obesity increases oxidative stress (oxidation) in the body. In turn, some scientists suspect that increased oxidation is the reason why obesity paves the way for neurodegenerative disorders like Alzheimer’s disease. Taking this into account, it might be helpful if potential drugs for treating obesity also had strong antioxidant functions.

In a global first, scientists have demonstrated that molecular robots are able to accomplish cargo delivery by employing a strategy of swarming, achieving a transport efficiency five times greater than that of single robots.

Swarm robotics is a new discipline, inspired by the cooperative behavior of living organisms, that focuses on the fabrication of robots and their utilization in swarms to accomplish complex tasks. A swarm is an orderly collective behavior of multiple individuals. Macro-scale swarm robots have been developed and employed for a variety of applications, such as transporting and accumulating cargo, forming shapes, and building complex structures.

A team of researchers, led by Dr. Mousumi Akter and Associate Professor Akira Kakugo from the Faculty of Science at Hokkaido University, has succeeded in developing the world’s first working micro-sized machines utilizing the advantages of swarming. The findings were published in the journal Science Robotics. The team included Assistant Professor Daisuke Inoue, Kyushu University; Professor Henry Hess, Columbia University; Professor Hiroyuki Asanuma, Nagoya University; and Professor Akinori Kuzuya, Kansai University.

The pancreas is a key metabolic regulator. When pancreatic beta cells cease producing enough insulin, blood sugar levels rise dangerously — a phenomenon known as hyperglycemia — thus triggering diabetes. After discovering that other mature pancreatic cells can adapt and partly compensate for the lack of insulin, a team from the University of Geneva (UNIGE) demonstrates that the stem cells from which beta cells are derived are only present during embryonic development. This discovery puts an end to a long-standing controversy about the hypothetical existence of adult pancreatic stem cells that would give rise to newly differentiated hormone-producing cells after birth. The scientists also succeeded in precisely defining the ‘identity card’ of pancreatic endocrine cells, which is a promising tool for the production of replacement insulin-secreting cells. These results can be read in Cell Reports and Nature Communications.

Diabetes is a common metabolic disease. It is characterised by a persistent hyperglycemia that occurs when pancreatic cells responsible for the production of insulin — the beta cells — are destroyed or are no longer able to produce this regulatory hormone in sufficient quantities. Since 2010, studies performed by the team of Pedro Herrera, a professor in the Department of Genetic Medicine and Development and in the Diabetes Centre at the UNIGE Faculty of Medicine, as well as at the Geneva Institute of Genetics and Genomics (iGE3), reveal that the other pancreatic endocrine cells — namely alpha, delta and gamma cells, which produce other hormones useful for the metabolic balance — can “learn” to produce insulin when beta cells are absent or defective. This phenomenon, observed in mice and humans, demonstrates the plasticity of pancreatic cells and paves the way to new therapeutic strategies.

3D printed ICs overcome semiconductor shortage and reinforce supply chains, leading to cheaper production costs, faster prototyping, and faster time to market.

As the size of microchip packages shrinks, semiconductor manufacturers are under pressure to improve lithography capabilities. For more than two decades, researchers have been working on 3D printed integrated circuits. Earlier attempts at 3D printing electronics used the proper technique but failed to reach the required levels of conductivity for a PCB, leaving complicated electrical circuits unusable. Over time though, these printing tools have gradually improved feature size resolution, yield, and variability in production.

● A broad variety of materials may be used to produce an assortment of printed electronics, including conductors, semiconductors, dielectrics, resistors, and other components. ● As has been the case in other areas, additive manufacturing is expected to result in more innovative products, lower costs, and faster production runs.