Lifeboat Foundation

The original 2017 transformer model was designed for natural language processing (NLP), where it achieved SOTA results. Its performance intrigued machine learning researchers, who have since successfully adapted the attention-based architecture to perception tasks in other modalities, such as the classification of images, video and audio. While transformers have shown their power and potential in these areas, achieving SOTA performance requires training a separate model for each task. Producing a single transformer model capable of processing multiple modalities and datasets and sharing its learnable parameters has thus emerged as an attractive research direction.

To this end, a team from Google Research, University of Cambridge and Alan Turing Institute has proposed PolyViT; a single transformer architecture co-trained on image, audio and video that is parameter-efficient and learns representations that generalize across multiple domains.

The PolyViT design is motivated by the idea that human perception is inherently multimodal and previous studies that have demonstrated transformers’ ability to operate on any modality that can be tokenized. PolyViT shares a single transformer encoder across different tasks and modalities, enabling up to a linear reduction in parameters with the number of tasks.

Future Human Community

There is a huge global effort to engineer a computer capable of harnessing the power of quantum physics to carry out computations of unprecedented complexity. While formidable technological obstacles still stand in the way of creating such a quantum computer, today’s early prototypes are still capable of remarkable feats.

For example, the creation of a new phase of matter called a “time crystal.” Just as a crystal’s structure repeats in space, a time crystal repeats in time and, importantly, does so infinitely and without any further input of energy—like a clock that runs forever without any batteries. The quest to realize this phase of matter has been a longstanding challenge in theory and experiment—one that has now finally come to fruition.

In research published Nov. 30 in Nature, a team of scientists from Stanford University, Google Quantum AI, the Max Planck Institute for Physics of Complex Systems and Oxford University detail their creation of a time crystal using Google’s Sycamore quantum computing hardware.

Disney’s AI research division has developed a hybrid method for movie-quality facial simulation, combining the strengths of facial neural rendering with the consistency of a CGI-based approach. The pending paper is titled Rendering with Style: Combining Traditional and Neural Approaches for High Quality Face Rendering, and is previewed in a new 10-minute video at the […].

Using artificial intelligence, the xenobots were able to replicate themselves using a process not seen in humans or any other animals.

The Indian edtech giant Byju’s keeps getting bigger, having raised more than $4.5 billion since it was founded 10 years ago. This month the company made clear its ambitious research agenda: to achieve the science-fiction dream of building next-generation teaching aids with artificial intelligence.

Specifically, the company announced a new research-and-development hub, with offices in Silicon Valley, London and Bangalore, that will work on applying the latest findings from artificial intelligence and machine learning to new edtech products. The new hub, called Byju’s Lab, will also work on “moonshots” of developing new forms of digital tutoring technology, said Dev Roy, chief innovation and learning officer for BYJU’s, in a recent interview with EdSurge.

“Edtech is one of the slowest adopters of AI so far, compared to some of the other industries out there,” Roy said. “Even in health care, what DeepMind has done with mapping the proteins of DNA—nobody’s doing that in the education sector.”

A team scoured the human proteome for antimicrobial molecules and found thousands, plus a surprise about how animals evolved to fight infections.

Forget losing your job to robots, Scientists have created robots that can reproduce. ‘Xenobots’ are capable of ‘self-replicating’ themselves. They are made up of stem cells taken from frogs. Astounded? Watch this report by Palki Sharma for the details.

#Gravitas #Robots #Xenobots.

Continue reading “Gravitas: These robots can produce babies | ‘Xenobots’ capable of ’self-replicating‘” »

When humans look at a scene, they see objects and the relationships between them. On top of your desk, there might be a laptop that is sitting to the left of a phone, which is in front of a computer monitor.

Many deep learning models struggle to see the world this way because they don’t understand the entangled relationships between individual objects. Without knowledge of these relationships, a robot designed to help someone in a kitchen would have difficulty following a command like “pick up the spatula that is to the left of the stove and place it on top of the cutting board.”

In an effort to solve this problem, MIT researchers have developed a model that understands the underlying relationships between objects in a scene. Their model represents individual relationships one at a time, then combines these representations to describe the overall scene. This enables the model to generate more accurate images from text descriptions, even when the scene includes several objects that are arranged in different relationships with one another.