Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 1,031

Preventing Colorectal Cancer With AI Technology

Patients who come to any Northwestern Medicine location for colonoscopies now have access to advanced artificial intelligence (AI) technology, which is improving the way gastroenterologists detect colon polyps and prevent colorectal cancer. According to new research by Northwestern Medicine, physicians who performed colonoscopies assisted by AI achieved a 13% increase in the detection and removal of colorectal polyps.

Computer-aided colonoscopies could reduce future colon cancer diagnoses by up to 39%.

“Most polyps do not become cancerous, but nearly all colorectal cancers begin as polyps,” says Rajesh N. Keswani, MD, MS, director of Endoscopy for Northwestern Memorial Hospital and director of quality for the Northwestern Medicine Digestive Health Center. “We want to detect them in their earliest stages and remove them to prevent future diagnoses of cancer. There’s nothing better than telling a patient that their decision to have a screening colonoscopy may have saved their life.”

Future AI algorithms have potential to learn like humans, say researchers

Memories can be as tricky to hold onto for machines as they can be for humans. To help understand why artificial agents develop holes in their own cognitive processes, electrical engineers at The Ohio State University have analyzed how much a process called “continual learning” impacts their overall performance.

Continual learning is when a computer is trained to continuously learn a sequence of tasks, using its accumulated knowledge from old tasks to better learn new tasks.

Yet one major hurdle scientists still need to overcome to achieve such heights is learning how to circumvent the machine learning equivalent of memory loss—a process which in AI agents is known as “catastrophic forgetting.” As are trained on one new task after another, they tend to lose the information gained from those previous tasks, an issue that could become problematic as society comes to rely on AI systems more and more, said Ness Shroff, an Ohio Eminent Scholar and professor of computer science and engineering at The Ohio State University.

AI-Powered NPCs: A Game-Changing FREE Demo

Once this is added to a VR open world, like low-fi, I may never come out!

The Unreal Engine combined with the power of OpenAI’s GPT has opened up a lot of possibilities for the future of video games. This demo created by Replica Studios allows us to directly interact with NPCs. It’s surprisingly good, and gives some insight into where things might be heading.

Download the demo from here: https://blog.replicastudios.com/smart-npc-plugin-release/

If you would like to help support my channel, please check my Patreon page: https://www.patreon.com/obsidianant.
PayPal Donation: https://www.paypal.com/donate/?cmd=_s-xclick&hosted_button_id=H5LJ5URGUB7ZQ
Any support is massively appreciated!

Music.

Continue reading “AI-Powered NPCs: A Game-Changing FREE Demo” | >

Falcon 9 puts on a show in the Californian skies

Last night at 9:09 p.m. PT (04:09 UTC), SpaceX successfully launched 15 V2 mini Starlink satellites to low Earth orbit from Space Launch Complex 4 East at Vandenberg Space Force Base, California.

Following a last-second aborted launch attempt the previous night, SpaceX teams performed checkouts of the Falcon 9 and determined another attempt to launch the next-gen satellites was good to go. During the previous attempt, the launch was held early in the countdown due to a “perceived leak” in the second stage, then eventually, the automated abort at T-minus 5 seconds.

Liftoff! pic.twitter.com/CzetVZFGbv

Apple is quietly working on ‘Apple GPT’ to rival OpenAI

It is believed that Apple is creating its large language model in competition with OpenAI’s GPT-3 and GPT-4, Google’s Bert and LaMDA, Meta’s LLaMa-1 and LLaMa-2, among others.

Whenever Apple launches a new product, its competitors, developers, and starry-eyed consumers take it very seriously. And with the veritable craze over artificial intelligence (AI) tools, it was only a matter of time before the iPhone creator came up with something explosive.

It is believed that Apple is creating its large language model (LLM) in competition with OpenAI’s GPT-3 and GPT-4, Google’s Bert and LaMDA, Meta’s LLaMa-1 and LLaMa-2, among others. It is also being learned that Apple has created a chatbot service… More.

Getty Images.

New dual-resolution technique opens door for faster drone exploration

Researchers from Carnegie Mellon University have developed a new technique that could lead to faster and more efficient drone exploration.

A team of researchers from Carnegie Mellon University has successfully developed a new dual-mapping technique that could help robots explore areas faster and more efficiently. By producing both a site’s high-and low-resolution map, this new technique enables robots to explore areas using only a fraction of the computing power typically needed for a similar task.

ROBOTICS INSTITUTE, CARNEGIE MELLON UNIVERSITY

More efficient exploration.

Meta announces ‘new and improved’ LLaMa 2 in partnership with Microsoft

Microsoft also has a collaboration with OpenAI, which mostly develops closed-source LLMs.

Meta has partnered with Microsoft to launch an open-source large language model like ChatGPT. Called LLaMa 2, it was trained on 40% more data in comparison to LLaMa 1, which Meta had launched in February.

What sets it apart from its competitors like OpenAI’s ChatGPT and Google’s Bard is that it’s open source.

By collaborating with Meta, which has always been a proponent of open-source models, Microsoft has its… More.

Instagram.

In a press release, Meta claims that LLaMa 2 has double the context length vs. LLaMa 1 and “outperforms” other LLMs like Falcon and MPT regarding reasoning, coding, proficiency, and knowledge tests.

Continue reading “Meta announces ‘new and improved’ LLaMa 2 in partnership with Microsoft” | >

Stanford researchers claim ChatGPT’s performance and accuracy has decreased over time

Researchers compared performance of OpenAI’s GPT-3.5 and GPT-4.

It seems that the honeymoon phase for large language models (LLMs), introduced in the rush to make inroads in the generative AI space, is over. According to a study by researchers at Stanford and UC Berkeley, the performance of OpenAI’s LLMs has decreased significantly over time.

The researchers wanted to determine if these LLMs were improving, as they can be updated based on data, user feedback, and design changes.

Robert Way/iStock.

According to a study by researchers at Stanford and UC Berkeley, the performance of OpenAI’s LLMs has decreased significantly over time.

/* */