Lifeboat Foundation

Researchers from LP3 Laboratory in France developed a light-based technique for local material processing anywhere in the three-dimensional space of semiconductor chips. The direct laser writing of new functionalities opens the possibility to exploit the sub-surface space for higher integration densities and extra functions.

Semiconductors remain the backbone material of the electronics integrated with modern devices such as cellphones, cars, robots and many other intelligent devices. Driven by the continuous need for miniaturized and powerful chips, the current semiconductor manufacturing technologies are facing increasing pressure.

The dominating manufacturing technology, lithography, has strong limitations when addressing these challenges, given its surface processing nature. For this reason, a solution to fabricating structures under the wafer surfaces would be highly desirable so that the full space inside the materials could be exploited.

😀

A Paris-based startup has created a genetically engineered houseplant that can literally clean the air within your home. The plant builds off the natural purifying properties that houseplants already offer. So, while it adds some color to whatever room you put it in, it’s also actively keeping the air cleaner than 30 air purifiers.

The company, called Neoplants, modified both a pothos plant as well as its root microbiome to pump the plant’s natural air-cleaning properties up quite a bit. Called Neo P1, the genetically engineered houseplant recently hit the market, and you can purchase it right now.

Continue reading “New genetically engineered houseplant cleans air as efficiently as 30 air purifiers” »

T ransplant medicine could take a giant leap forward if donor organs could soak up oxygen for longer and decay delayed. A technology called OrganEx, described in Nature from a team at Yale, promises to do just that. The researchers stopped the hearts of pigs and an hour later used OrganEx, then cataloged the return of bodily functions. The new approach far exceeded the ability of existing technology to prolong organ viability.

Pigs have long been a popular animal model of human disease because they are about our size and their hearts and blood vessels are quite similar. They have also had fictional roles in medicine.

In the Twilight Zone episode Eye of the Beholder, Janet Tyler has undergone multiple procedures to replace the “pitiful twisted lump of flesh” that is her face with something more acceptable. At the end, as the bandages are slowly unrolled from yet another failed procedure, we see that she naturally looks like us, considered hideous in her world where most people, including the nurse and doctor, have pig faces. Janet and others like her are sent to live among themselves.

Artificial intelligence has helped design an invisibility cloak. The cloak could hide communication devices from detectors that use microwaves or infrared light.

Quantum mechanics, the theory which rules the microworld of atoms and particles, certainly has the X factor.

Unlike many other areas of physics, it is bizarre and counter-intuitive, which makes it dazzling and intriguing.

When the 2022 Nobel prize in physics was awarded to Alain Aspect, John Clauser, and Anton Zeilinger for research shedding light on quantum mechanics, it sparked excitement and discussion.

How scientists are using virtual reality to create avatars, chatbots and even eternal digital entities.

Artificial intelligence is rapidly transforming all sectors of our society. Whether we realize it or not, every time we do a Google search or ask Siri a question, we’re using AI.

For better or worse, the same is true about the very character of warfare. This is the reason why the Department of Defense – like its counterparts in China and Russia– is investing billions of dollars to develop and integrate AI into defense systems. It’s also the reason why DoD is now embracing initiatives that envision future technologies, including the next phase of AI – artificial general intelligence.

AGI is the ability of an intelligent agent to understand or learn any intellectual task in the same way that humans do. Unlike AI which relies on ever-expanding datasets to perform more complex tasks, AGI will exhibit the same attributes as those associated with the human brain, including common sense, background knowledge, transfer learning, abstraction, and causality. Of particular interest is the human ability to generalize from scanty or incomplete input.

The magic of the Internet.

Large models have improved performance on a wide range of modern computer vision and, in particular, natural language processing problems. However, issuing patches to adjust model behavior after deployment is a significant challenge in deploying and maintaining such models. Because of the distributed nature of the model’s representations, when a neural network produces an undesirable output, making a localized update to correct its behavior for a single or small number of inputs is difficult. A large language model trained in 2019 might assign a higher probability to Theresa May than Boris Johnson when prompted. Who is the Prime Minister of the United Kingdom?

An ideal model editing procedure would be able to quickly update the model parameters to increase the relative likelihood of Boris Johnson while not affecting the model output for unrelated inputs. This procedure would yield edits with reliability, successfully changing the model’s work on the problematic input (e.g., Who is the Prime Minister of the United Kingdom?); locality, affecting the model’s output for unrelated inputs (e.g., What sports team does Messi play for?); and generality, generating the correct output for inputs related to the edit input (e.g., Who is the Prime Minister of the United Kingdom?). Making such edits is as simple as fine-tuning with a new label on the single example to be corrected. However, fine-tuning on a single sample tends to overfit, even when the distance between the pre-and post-fine-tuning parameters is limited.

Overfitting causes both locality and generality failures. While fine-tuning the edit example and ongoing training on the training set improves locality, their experiments show that it still needs more generality. Furthermore, it necessitates continuous access to the entire training set during testing and is more computationally demanding. Recent research has looked into methods for learning to make model edits as an alternative. Researchers present a bi-level meta-learning objective for determining a model initialization for which standard fine-tuning on a single edit example yields valuable modifications.