Goto

Collaborating Authors

Obituary


Technoblade, 'Minecraft' YouTube creator, dies aged 23

Mashable

Technoblade, top Minecraft YouTuber, has died aged 23 following a battle with cancer. The creator's father posted a moving video on YouTube on Thursday titled "so long nerds," with a message written by Technoblade himself. The video had been planned with his family over his final months, but was written by Technoblade about eight hours before he died, his father said. In the message, he revealed his real name: Alex. I am dead," he wrote in the message, read by his father. "Thank you all for supporting my content over the years.


Megvii Chief Scientist Sun Jian dies. Leaves behind notable AI innovations on 2 continents

#artificialintelligence

A bright mind in AI and biometrics, Megvii innovator Sun Jian has died aged 45 after a sudden illness. No more information has been released about Jian's final days. Sun had researched computer vision and computational photography for 13 years at Microsoft, leaving in 2016 for Megvii, then an AI and facial recognition startup. He became that firm's chief scientist and managing director of research, according to the South China Morning Post. At Megvii, Sun managed the development of Brain, which became the company's core AI productivity platform.


'World's first cyborg' dies aged 64: British doctor passes away peacefully

Daily Mail - Science & tech

A British scientist who became the'world's first full cyborg' has passed away at the age of 64. Peter Scott-Morgan decided to challenge what it meant to be human when he refused to accept his fate following a diagnosis of motor neurone disease in 2017. He said he wanted to push the boundaries of what science can achieve so decided to extend his life and become fully robotic. His family confirmed the news of his passing on his Twitter account this morning. 'To Peter's amazing rebel supporters: With a broken heart, I'm letting you all know that Peter passed peacefully surrounded by his family, and those closest to him,' they wrote.


Top scientist at Chinese AI giant Megvii dead at 45

#artificialintelligence

A staff member speaks in front of a display demonstrating the facial recognition system of Chinese artificial intelligence firm Megvii during an organised media tour at the Zhongguancun National Innovation Demonstration Zone Exhibition Center in Beijing on February 10, 2022.


Artificial intelligence spots type 1 diabetes in children earlier

#artificialintelligence

A predictive tool using artificial intelligence could provide hope for earlier diagnosis of type 1 diabetes in children across the UK, reducing the risk of potentially fatal diabetic ketoacidosis (DKA), early research presented at the Diabetes UK Professional Conference 2022 has revealed. Type 1 diabetes is a serious auto-immune condition that cannot yet be prevented, and the gradual destruction of insulin-making beta cells can start months or even years before being diagnosed. Symptoms usually start to appear much closer to diagnosis. Early diagnosis and awareness of the signs and symptoms of diabetes are crucial to ensure that both children and adults who develop it do not become critically ill. A quarter of children and young people (25%1) aren't diagnosed with type 1 diabetes until they are in DKA2, a life-threatening condition that can lead to coma or even death.


How to remember the Japanese incarceration, 80 years later

Los Angeles Times

Akemi Leung knew her grandfather had been incarcerated at Heart Mountain in Wyoming during World War II. But he never spoke much about it. Only when she read and watched a video of his testimony at a congressional commission hearing did she learn more about what he suffered as one of more than 120,000 Americans of Japanese ancestry forced to leave their homes and live in concentration camps. "I just knew him to be a quiet person who liked to observe more than talk," Leung said. "Seeing the testimony helped illustrate how he was a leader."


Elon Musk-owned Neuralink confirms monkeys died during tests but rejects abuse claim

Daily Mail - Science & tech

Elon Musk's brain-chip firm Neuralink has admitted monkeys died during tests, but denied claims of animal abuse put forward by an animal rights group. The biotech firm is developing a brain-computer interface, that it claims could one day make humans hyper-intelligent, and allow paralyzed people to walk again. Last week the Physicians Committee for Responsible Medicine (PCRM) lodged a complaint with the US Department of Agriculture, alleging several counts of animal abuse between 2017 and 2020, involving test monkeys owned by Neuralink. They claimed the macaque monkeys, housed at a University of California Davis research facility, were subject to experiments that amounted to torture, with evidence of rashes, self-mutilation and brain hemorrhages seen in documentation. Neuralink has hit back at the claims of abuse, calling out the PCRM as a group that oppose any use of animals in research.


Farewell Douglas Trumbull, visual effects pioneer

Engadget

If you've watched a classic, landmark sci-fi movie and you were blown away by the quality and realism of its effects, then there's a good chance Douglas Trumbull's name is in the credits. The VFX pioneer, who passed away on February 8th, 2022, has worked on key films in the sci-fi canon. Even a short version of his resume would have to include 2001: A Space Odyssey, Close Encounters of the Third Kind, Star Trek: The Motion Picture, Blade Runner and Silent Running. To have worked on one of those in your lifetime would have been a big deal, but to have contributed to all of them speaks to just how much work Trumbull did to push the artform forward. Trumbull was the son of an artist and engineer, Donald Trumbull, who worked on VFX for The Wizard of Oz.


Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents

arXiv.org Artificial Intelligence

Can world knowledge learned by large language models (LLMs) be used to act in interactive environments? In this paper, we investigate the possibility of grounding high-level tasks, expressed in natural language (e.g. "make breakfast"), to a chosen set of actionable steps (e.g. "open fridge"). While prior work focused on learning from explicit step-by-step examples of how to act, we surprisingly find that if pre-trained LMs are large enough and prompted appropriately, they can effectively decompose high-level tasks into low-level plans without any further training. However, the plans produced naively by LLMs often cannot map precisely to admissible actions. We propose a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions. Our evaluation in the recent VirtualHome environment shows that the resulting method substantially improves executability over the LLM baseline. The conducted human evaluation reveals a trade-off between executability and correctness but shows a promising sign towards extracting actionable knowledge from language models. Website at https://huangwl18.github.io/language-planner


The Introspective Agent: Interdependence of Strategy, Physiology, and Sensing for Embodied Agents

arXiv.org Artificial Intelligence

The last few years have witnessed substantial progress in the field of embodied AI where artificial agents, mirroring biological counterparts, are now able to learn from interaction to accomplish complex tasks. Despite this success, biological organisms still hold one large advantage over these simulated agents: adaptation. While both living and simulated agents make decisions to achieve goals (strategy), biological organisms have evolved to understand their environment (sensing) and respond to it (physiology). The net gain of these factors depends on the environment, and organisms have adapted accordingly. For example, in a low vision aquatic environment some fish have evolved specific neurons which offer a predictable, but incredibly rapid, strategy to escape from predators. Mammals have lost these reactive systems, but they have a much larger fields of view and brain circuitry capable of understanding many future possibilities. While traditional embodied agents manipulate an environment to best achieve a goal, we argue for an introspective agent, which considers its own abilities in the context of its environment. We show that different environments yield vastly different optimal designs, and increasing long-term planning is often far less beneficial than other improvements, such as increased physical ability. We present these findings to broaden the definition of improvement in embodied AI passed increasingly complex models. Just as in nature, we hope to reframe strategy as one tool, among many, to succeed in an environment. Code is available at: https://github.com/sarahpratt/introspective.