Lifeboat Foundation

Kawasaki has shoehorned the supercharged 1,000cc engine from its wild H2R hyperbike into a heavy-lift autonomous cargo helicopter, and has now demonstrated a robotic system for loading and unloading it without exposing humans to those big blades.

The K-Racer X1 is a beast of a drone, roughly the size of a small car. It rises vertically on a helicopter-style top rotor, but where there’s normally a tail rotor to balance out torque, this machine uses two forward-facing props mounted at the end of stubby wings. These props double as forward propulsion, with the wings providing some lift.

Continue reading “Kawasaki’s supercharged cargo copter tests robotic ground crew” »

Poor Artificial Intelligence (AI). For years, it has had to sit there (like a dormant Skynet) listening to its existence being debated, without getting to have a say. A recent debate held at the University of Oxford tried to put that right by including an AI participant in a debate on the topic of whether AI can ever be ethical.

The debate involved human participants, as well as the Megatron Transformer, an AI created by the Applied Deep Research team at computer-chip maker Nvidia. The Megatron has been trained on a dataset called “the pile”, which includes the whole of Wikipedia, 63 million English news articles, and 38 gigabytes of Reddit conversations — more than enough to break the mind of any human forced to do likewise.

“In other words, the Megatron is trained on more written material than any of us could reasonably expect to digest in a lifetime,” Oxford’s Professor Andrew Stephen wrote in a piece on the debate published in The Conversation. “After such extensive research, it forms its own views.”

Billionaire Mukesh Ambani’s Reliance Industries Ltd. is buying an Indian robotics startup as the conglomerate scales up automation across its businesses, from e-commerce to new energy.

This is lecture 3 of course 6.S094: Deep Learning for Self-Driving Cars taught in Winter 2017. This lecture introduces computer vision, convolutional neural networks, and end-to-end learning of the driving task.

INFO:
Slides: http://bit.ly/2HdXYvf.
Website: https://deeplearning.mit.edu.
GitHub: https://github.com/lexfridman/mit-deep-learning.
Playlist: https://goo.gl/SLCb1y.

Continue reading “MIT 6.S094: Convolutional Neural Networks for End-to-End Learning of the Driving Task” »

Part of the End-to-End Machine Learning School Course 193, How Neural Networks Work at https://e2eml.school/193

Slides: https://docs.google.com/presentation/d/1R-DnrghbU36jO8X4scbr…sp=sharing.
machine learning blog: https://brohrer.github.io/blog.html

SEOUL — Using a high-performance artificial intelligence (AI) chip, South Korean researchers have established a system that can accelerate the process of learning data and yielding results. The system capable of performing five thousand trillion operations per second is ideal for autonomous vehicles and AI servers because its chipset is about the size of a coin.

One possible application of interest to researchers is bone healing. The idea is that the soft material, powered by the electroactive polymer, will be able to maneuver in spaces of complicated bone fractures and expand. When the material hardens, it can form the basis for building new bones. In their study, the researchers demonstrate that the material can wrap itself around chicken bones, and the artificial bone that develops later grows along with the animal’s bone. The developed biohybrid variable-stiffness actuators can be used in soft (micro-)robotics and as potential tools for bone repair or bone tissue engineering.

“By controlling how the material turns, we can make the microrobot move in different ways, and also affect how the material unfurls in broken bones. We can embed these movements into the material’s structure, making complex programs for steering these robots unnecessary”, says Edwin Jager.

Tech companies continued to draw criticism for their roles in political and social scandals, most notably when whisteblower and former Facebook employee Frances Haugen testified to lawmakers. Undeterred, Facebook rebranded itself Meta and said it would now focus on building the metaverse. Twitter CEO Jack Dorsey stepped down and likewise changed the name of his company Square to Block in a not-so-subtle nod to the blockchain.

Meanwhile, volatile cryptocurrencies set new records, their prices jumping and crashing on a tweet. NFTs, a once-obscure type of cryptoasset, went on an eye-watering tear as redditors pushed meme stocks skyward. It was also the year of ever-bigger AI. Machine learning models surpassed a trillion parameters, designed computer chips, and tackled practical problems in biology, math, and chemistry. Elsewhere, billionaires went to space, regular folks bought 3D printed houses, fusion power attracted billions in investment, gene editing trials hit their stride, and “flying car” companies hit the New York Stock Exchange.

For this year’s list of fascinating stories in tech and science, we sifted our Saturday posts and selected articles that looked back to where it all began, glanced ahead to what’s coming, or otherwise stood out from the chatter to stand the test of time.

Over the last several decades, I’ve witnessed a lot of change in the fields of machine learning (ML) and computer science. Early approaches, which often fell short, eventually gave rise to modern approaches that have been very successful. Following that long-arc pattern of progress, I think we’ll see a number of exciting advances over the next several years, advances that will ultimately benefit the lives of billions of people with greater impact than ever before. In this post, I’ll highlight five areas where ML is poised to have such impact. For each, I’ll discuss related research (mostly from 2021) and the directions and progress we’ll likely see in the next few years.