Lifeboat Foundation

A group of scientists from the University College London has developed an artificial intelligence (AI) algorithm that can detect drug-resistant focal cortical dysplasia (FCD), a subtle anomaly in the brain that leads to epileptic seizures. This is a promising step for scientists toward detecting and curing epilepsy in its early stages.

To develop the algorithm, the Multicentre Epilepsy Lesion Detection project (MELD) gathered more than 1,000 patients’ MRI scans from 22 international epilepsy centers, which reports where anomalies are in cases of drug-resistant focal cortical dysplasia (FCD), a major reason behind epilepsy.

Functioning in curved space, the robot heralds new space locomotive technology possibilities without the use of propellants.

A robot engineered at Georgia Institute of Technology (Georgia Tech) has done the unthinkable and flouted a steadfast law of motion, suggesting that new laws need to be defined. Such new principles may have applications in new forms of locomotion without propellants.

We’ve all seen the hilarious slapstick gag where the unwitting individual steps on a banana peel, landing comically on their rump. It may not seem like it, but the quip relies on the fact that human locomotion, like all locomotion, is based on Newton’s third law of motion.

Continue reading “Curved space robot defies known laws of physics” »

Circa 2018

This is a talk by Stephen Wolfram for MIT course 6.S099: Artificial General Intelligence. This class is free and open to everyone. Our goal is to take an engineering approach to exploring possible paths toward building human-level intelligence for a better world.

Continue reading “Stephen Wolfram: Computational Universe | MIT 6.S099: Artificial General Intelligence (AGI)” »

But what I find even more interesting is that as metaverse tools like Nvidia’s Omniverse become more consumer friendly, the ability to use AI and human digital twins will enable us to create our own worlds where we dictate the rules and where our AI-driven digital twins will emulate real people and animals.

At that point, I expect we’ll need to learn what it means to be gods of the worlds we create, and I doubt we are anywhere near ready, both in terms of the addictive nature of such products and how to create these metaverse virtual worlds in ways that can become the basis for our own digital immortality.

Continue reading “The Metaverse Future: Are You Ready To Become a God?” »

The quadrupedal robots are well suited for repetitive tasks.

Mankind’s new best friend is coming to the U.S. Space Force.

The Space Force has conducted a demonstration using dog-like quadruped unmanned ground vehicles (Q-UGVs) for security patrols and other repetitive tasks. The demonstration used at least two Vision 60 Q-UGVs, or “robot dogs,” built by Ghost Robotics and took place at Cape Canaveral Space Force Station on July 27 and 28.

Register FREE. Streamed online.

In 1903, the Wright brothers invented the first successful airplane. By 1914, just over a decade after its successful test, aircraft would be used in combat in World War I, with capabilities including reconnaissance, bombing and aerial combat. This has been categorized by most historians as a revolution in military affairs. The battlefield, which previously included land and sea, now included the sky, permanently altering the way wars are fought. With the new technology came new strategy, policy, tactics, procedures and formations.

Twenty years ago, unmanned aircraft systems (UASs) were much less prevalent and capable. Today, their threat potential and risk profile have increased significantly. UASs are becoming increasingly more affordable and capable, with improved optics, greater speed, longer range and increased lethality.

The U.S. has long been a proponent of utilizing unmanned aircraft systems, with the MQ-9 Reaper and MQ-1 Predator excelling in combat operations, and smaller squad-based UASs being fielded, such as the RQ-11 Raven and the Switchblade. While the optimization of friendly UAS capability can yield great results on the battlefield, adversarial use of unmanned aircraft systems can be devastating.

Around the same time, neuroscientists developed the first computational models of the primate visual system, using neural networks like AlexNet and its successors. The union looked promising: When monkeys and artificial neural nets were shown the same images, for example, the activity of the real neurons and the artificial neurons showed an intriguing correspondence. Artificial models of hearing and odor detection followed.

But as the field progressed, researchers realized the limitations of supervised training. For instance, in 2017, Leon Gatys, a computer scientist then at the University of Tübingen in Germany, and his colleagues took an image of a Ford Model T, then overlaid a leopard skin pattern across the photo, generating a bizarre but easily recognizable image. A leading artificial neural network correctly classified the original image as a Model T, but considered the modified image a leopard. It had fixated on the texture and had no understanding of the shape of a car (or a leopard, for that matter).

Self-supervised learning strategies are designed to avoid such problems. In this approach, humans don’t label the data. Rather, “the labels come from the data itself,” said Friedemann Zenke, a computational neuroscientist at the Friedrich Miescher Institute for Biomedical Research in Basel, Switzerland. Self-supervised algorithms essentially create gaps in the data and ask the neural network to fill in the blanks. In a so-called large language model, for instance, the training algorithm will show the neural network the first few words of a sentence and ask it to predict the next word. When trained with a massive corpus of text gleaned from the internet, the model appears to learn the syntactic structure of the language, demonstrating impressive linguistic ability — all without external labels or supervision.

In this award-winning AutoML conference paper, Amazon Web Services and ETH Zürich scientists present a new way to decide when to terminate Bayesian optimization… See more.

Bayesian optimization (BO) is a widely popular approach for the hyperparameter optimization (HPO) in machine learning. At its core, BO iteratively evaluates promising configurations until a user-defined budget, such as wall-clock time or number of iterations, is exhausted. While the final performance after tuning heavily depends on the provided budget, it is hard to pre-specify an optimal value in advance. In this work, we propose an effective and intuitive termination criterion for BO that automatically stops the procedure if it is sufficiently close to the global optimum. Our key insight is that the discrepancy between the true objective (predictive performance on test data) and the computable target (validation performance) suggests stopping once the sub-optimality in optimizing the target is dominated by the statistical estimation error. Across an extensive range of real-world HPO problems and baselines, we show that our termination criterion achieves a better trade-off between the test performance and optimization time. Additionally, we find that overfitting may occur in the context of HPO, which is arguably an overlooked problem in the literature, and show how our termination criterion helps to mitigate this phenomenon on both small and large datasets.