Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 1112

Artificial-intelligence systems are increasingly limited by the hardware used to implement them. Now comes a new superconducting photonic circuit that mimics the links between brain cells—burning just 0.3 percent of the energy of its human counterparts while operating some 30,000 times as fast.

In artificial neural networks, components called neurons are fed data and cooperate to solve a problem, such as recognizing faces. The neural net repeatedly adjusts the synapses—the links between its neurons—and determines whether the resulting patterns of behavior are better at finding a solution. Over time, the network discovers which patterns are best at computing results. It then adopts these patterns as defaults, mimicking the process of learning in the human brain.

Read more | >

“Machine learning provides a way of providing almost human-like intuition to huge data sets. One valuable application is for tasks where it’s difficult to write a specific algorithm to search for something—human faces, for instance, or perhaps ” something strange,” wrote astrophysicist and Director of the Penn State University Extraterrestrial Intelligence Center, Jason Wright in an email to The Daily Galaxy. ” In this case, you can train a machine-learning algorithm to recognize certain things you expect to see in a data set,” Wright explains, ” and ask it for things that don’t fit those expectations, or perhaps that match your expectations of a technosignature.

Crowdsourcing Alien Structures

For instance,’ Wright notes, theoretical physicist Paul Davies has suggested crowdsourcing the task of looking for alien structures or artifacts on the Moon by posting imaging data on a site like Zooniverse and looking for anomalies. Some researchers (led by Daniel Angerhausen) have instead trained machine-learning algorithms to recognize common terrain features, and report back things it doesn’t recognize, essentially automating that task. Sure enough, the algorithm can identify real signs of technology on the Moon—like the Apollo landing sites!

Read more | >

What’s your favorite ice cream flavor? You might say vanilla or chocolate, and if I asked why, you’d probably say it’s because it tastes good. But why does it taste good, and why do you still want to try other flavors sometimes? Rarely do we ever question the basic decisions we make in our everyday lives, but if we did, we might realize that we can’t pinpoint the exact reasons for our preferences, emotions, and desires at any given moment.

Most AI systems are black box models, which are systems that are viewed only in terms of their inputs and outputs. Scientists do not attempt to decipher the “black box,” or the opaque processes that the system undertakes, as long as they receive the outputs they are looking for.

Read more | >

Up until recently, artificial intelligence was unable to perform such creative-looking tasks.

But all of that is beginning to change thanks to AI Sketch software like DreamStudio, Dall-E 2, and Stable Diffusion, which take a few keywords via a text interface to generate an image in a process known as “generative AI.”

Generative AI is trained on sets of images, which are sourced from the internet. The machine can then learn the differences between people, places, and things and generate its own images from any text it receives.

The more data sets the AI can draw from, the more accurate and creative the results.

Continue reading “AI Just Got More Creative, and Filmmakers Shouldn’t Be Afraid” | >

Today, Replit announced Ghostwriter, an AI-powered programming assistant that can make suggestions to make coding easier. It works within Replit’s online development environment and resembles GitHub Copilot’s ability to recognize and compose code in various programming languages to accelerate the development process.

According to Replit, Ghostwriter works by using a large language model trained on millions of lines of publicly available code. This baked-in data allows Ghostwriter to make suggestions based on what you’ve already typed while programming in Replit’s IDE. When you see a suggestion you like, you can “autocomplete” the code by pressing the Tab key.

Read more | >

Greg Brockman, President and Co-Founder of @OpenAI, joins Alexandr Wang, CEO and Founder of Scale, to discuss the role of foundation models like GPT-3 and DALL·E 2 in research and in the enterprise. Foundation models make it possible to replace task-specific models with those that are generalized in nature and can be used for different tasks with minimal fine-tuning.

In January 2021, OpenAI introduced DALL·E, a text-to-image generation program. One year later, it introduced DALL·E 2, which generates more realistic, accurate, lower-latency images with four times greater resolution than its predecessor. At the same time, it released InstructGPT, a large language model (LLM) explicitly designed to follow instructions. InstructGPT makes it practical to leverage the OpenAI API to revise existing content, such as rewriting a paragraph of text or refactoring code.

Before creating OpenAI, Brockman was the CTO of Stripe, which he helped build from four to 250 employees. Watch this talk to learn how foundation models can help businesses benefit from applications that they can create more quickly than with past generations of AI tools.

Read more | >

Imagine the booming chords from a pipe organ echoing through the cavernous sanctuary of a massive, stone cathedral.

The a cathedral-goer will hear is affected by many factors, including the location of the organ, where the listener is standing, whether any columns, pews, or other obstacles stand between them, what the walls are made of, the locations of windows or doorways, etc. Hearing a sound can help someone envision their environment.

Researchers at MIT and the MIT-IBM Watson AI Lab are exploring the use of spatial acoustic information to help machines better envision their environments, too. They developed a that can capture how any sound in a room will propagate through the space, enabling the model to simulate what a listener would hear at different locations.

Read more | >

The project, known as DAF-MIT AI Accelerator, selected a pilot out of over 1,400 applicants.

The United States Air Force (DAF) and Massachusetts Institute of Technology (MIT) commissioned their lead AI pilot — a training program that uses artificial intelligence — in October 2022. The project utilizes the expertise at MIT and the Department of Air Force to research the potential of applying AI algorithms to advance the DAF and security.

The military department and the university created an artificial intelligence project called the Department of the Air Force-Massachusetts Institute of Technology Artificial Intelligence Accelerator (DAF-MIT AI Accelerator).

Full Story:

Continue reading “US Air Force and MIT commission a lead AI pilot for their innovative project” | >

Software upgrades could help resolve the issue.

A collaboration of researchers from the U.S. and Japan has demonstrated that a laser attack could be used to blind autonomous cars and delete pedestrians from their view, endangering those in its path, according to a press release.

Autonomous or self-driving cars rely on a spinning type of radar system called LIDAR that helps the vehicle sense its surroundings. Short for Light Detection and Ranging, the system emits laser lights and then captures its reflections to determine the distances between itself and the obstacles in its path.

Most advanced autonomous cars today rely on this system to steer through obstacles in their path. However, the collaboration of researchers from the University of Florida, the University of Michigan, and the University of Electro-Communications in Japan showed the system can be tricked with a fairly basic laser setup.

Continue reading “Researchers use lasers to trick autonomous cars and remove pedestrians from view” | >