Lifeboat Foundation

Convolutional neural networks used to be the typical network architecture in modern computer vision systems. Transformers with attention mechanisms were recently presented for visual tasks, and they performed well. Convolutions and self-attention aren’t required for MLP-based (multi-layer perceptron) vision models to perform properly. As a result of these advancements, vision models are reaching new heights.

The input image is handled differently by different networks. In Euclidean space, image data is commonly represented as a regular grid of pixels. On the image, CNNs apply a sliding window and introduce shift-invariance and locality. The MLP vision transformer, which was released recently, treats the image as a series of patches.

Recognizing the items in an image is one of the most basic tasks of computer vision. The typically utilized grid or sequence structures in prior networks like ResNet and ViT are redundant and inflexible to process because the objects are usually not quadrate whose shape is irregular. An item can be thought of as a collection of parts, such as a human’s head, upper torso, arms, and legs. These sections are organically connected by joints, forming a graph structure. Furthermore, a graph is a generic data structure, with grids and sequences being special cases of graphs. Visual perception is more flexible and effective when an image is viewed as a graph.

They can have abilities their creators did not foresee.

Huge “foundation models” are turbo — charging AI progress.

They can have abilities their creators did not foresee.

Machine learning can get a boost from quantum physics.

On certain types of machine learning tasks, quantum computers have an exponential advantage over standard computation, scientists report in the June 10 Science. The researchers proved that, according to quantum math, the advantage applies when using machine learning to understand quantum systems. And the team showed that the advantage holds up in real-world tests.

“People are very excited about the potential of using quantum technology to improve our learning ability,” says theoretical physicist and computer scientist Hsin-Yuan Huang of Caltech. But it wasn’t entirely clear if machine learning could benefit from quantum physics in practice.

Solving the 3D structure at near atomic level resolution, one of the world’s hardest, giant jigsaw puzzles—the nuclear pore complex—the largest molecular machine in human cells, with structure-based AI prediction @ScienceMagazine

We’ve been hearing a lot about synthetic skins designed for robotic hands, which would give the devices more human-like qualities. Well, scientists in Japan have gone a step further, by covering a robotic finger in a self-healing skin made from live human cells.

Led by Prof. Shoji Takeuchi, a team at the University of Tokyo started by building an articulated motor-driven robotic finger, capable of bending and straightening like its human counterpart. That finger was then submerged in a cylinder filled with a solution made up of collagen and human dermal fibroblast cells – these are the main components of our skin’s connective tissues.

Due to its natural properties, that solution shrank and conformed to the contours of the finger, forming a seamless hydrogel coating. Next, the scientists added a layer of human epidermal keratinocyte cells, which constitute 90 percent of our epidermis (the outermost layer of skin). These formed a moisture-retaining/water-resistant barrier on top of the gel, and gave the finger a more natural texture.

Numerous activities, including construction and demolition, mining and industrial activities, cooking and gardening, and others, generate a substantial amount of garbage. The amount of waste generated is directly proportional to consumption and production patterns.

In most cases, waste formation is the result of inefficient material utilization. Trends in the number, composition and impacts of these materials provide insight into the nation’s efficiency in using (and reusing) materials and resources. It also provides a better understanding of the effects of waste on human health and the environment.

According to surveys, 92 million tonnes of cloth are dumped as garbage each year worldwide. Estimates predict that this figure will likely exceed 130 million tonnes by 2030. When 200 tonnes of water used to make a single tonne of fabric is considered, it becomes clear that the end-to-end processes of the garment industry are severe threats to environmental initiatives.

In the hopes of one day building super realistic cyborgs, researchers built a robotic finger that wears living human skin.