Lifeboat Foundation

In the case of artificial intelligence, we have a problem. There is no clear, settled definition of natural intelligence. If we are not sure what the natural thing is, how can we know what the artificial thing ought to be?

In fact, I want to claim that intelligence is not a thing at all. It is an ongoing process. It is like science. You should not think of science as a body of absolute truth. Instead, think of the scientific method as a way of pursuing truth.

One should resist the temptation to think of intelligence as a huge lump of knowledge that an entity possesses. Memorizing the encyclopedia does not constitute intelligence.

The challenges of testing whether AI has rights.

Furthermore, the experimental values are introduced to correct the adsorption isotherms. For example, Fig. 3b shows the Langmuir adsorption isotherm obtained by fitting both the predicted and experimental adsorption data. While we use simulated datasets to address data scarcity, we can also properly introduce experimental values to correct adsorption isotherms, which helps a more quantitative prediction of adsorption performance at high-pressure where the gas-gas interaction becomes more significant. In Fig. 3b, one can observe that the corrected adsorption isotherms have a strong correlation with experimental adsorption capacity to some extent. The results exhibit that Uni-MOF not only has the ability to screen the adsorption performance of the same gas in different materials but also can accurately screen the adsorption performance of different gases in the same material (Fig. 3c, d) or at different temperatures (Fig. 3e, f).

In the foreseeable future, the intersection of Artificial Intelligence (AI) and materials science will necessitate the resolution of practical and scientific issues. Nonetheless, the attainment of process implementation by AI in the realm of machine learning techniques that entail copious amounts of data remains a formidable challenge, given the dearth of experimental data and the diverse array of synthetic technology and characterization conditions implicated. Our research has made a significant stride in materials science by incorporating operating conditions into the Uni-MOF framework to ensure data adequacy and enable screening functions that are consistent with experimental findings.

In order to showcase the predictive capabilities of Uni-MOF with regard to cross-system properties, five materials were randomly selected from each of the six systems (carbon-dioxide at 298 K, methane at 298 K, krypton at 273 K, xenon at 273 K, nitrogen at 77 K and argon at 87 K) contained in databases hMOF_MOFX_DB and CoRE_MOFX_DB, which have been thoroughly sampled in terms of temperature and pressure. The predicted and simulated values of gas adsorption uptake at varying pressures were then compared, with the results presented in Fig. 4a–f. Adsorption isotherms fitting from both Uni-MOF predictions and simulated values would artificially reduce visual errors. In order to eliminate data bias, adsorption isotherms in all cases were obtained only by simulated values. It is evident that, due to the fact that the adsorption isotherms were obtained purely through simulated values, the predicted values of adsorption uptake generated by Uni-MOF for the hMOF_MOFX_DB and CoRE_MOFX_DB databases align closely with the simulated values across all cases. This finding is further supported by the high prediction accuracy demonstrated in Fig. 2a, b.

A team of NUS researchers led by Associate Professor Lu Jiong from the Department of Chemistry and Institute for Functional Intelligent Materials, together with their international collaborators, have developed a novel concept of a chemist-intuited atomic robotic probe (CARP).

This innovation, which uses artificial intelligence (AI) to mimic the decision-making process of chemists, enables the manufacturing of quantum materials with unrivaled intelligence and precision for future quantum technology applications such as data storage and quantum computing.

Open-shell magnetic nanographene is a type of carbon-based quantum material that possesses key electronic and magnetic properties that are important for developing extremely fast electronic devices at the molecular level, or creating quantum bits, the building blocks of quantum computers. The processes used to develop such materials have progressed over the years due the discovery of a new type of solid-phase chemical reaction known as on-surface synthesis.

She continued: “We’ve certainly had more opportunities to target in the last 60 to 90 days,” adding the US is currently looking for “an awful lot” of rocket launchers in the region.

Moore’s comments provide some of the strongest evidence to date that the US military is using AI targeting systems to identify potential strike areas. She noted that even after Google walked away from the project, experimenting has continued with drone or satellite imagery.

Based at Central Command, or Centcom headquarters in Tampa, Florida, Moore revealed that US forces in the Middle East have been testing AI targeting systems using a combination of satellites and other data sources and conducted exercises over the past year with the technology.

“There’s a connection between the shape of the ice shell and the temperature in the ocean,” said Dr. Britney Schmidt. “This is a new way to get more insight from ice shell measurements that we hope to be able to get for Europa and other worlds.”

While Earth remains the only known world with bodies of liquid water on its surface, there are a myriad of worlds within our own solar system that have liquid water oceans beneath thick surfaces of ice. But what is the temperature of those interior oceans, and could the thickness of its ice shell determine it? This is what a recent study published in Journal of Geophysical Research Planets hopes to address as a team of researchers led by Cornell University investigated how a process called “ice pumping” could determine the temperature of the interior ocean underneath thick icy shells, also known as ice-ocean interaction. This study holds the potential to help researchers better understand the conditions for finding life beyond Earth with a focus on Jupiter’s moon, Europa, and Saturn’s moon, Enceladus.

“If we can measure the thickness variation across these ice shells, then we’re able to get temperature constraints on the oceans, which there’s really no other way yet to do without drilling into them,” said Dr. Britney Schmidt, who is an Associate Professor of Astronomy & Earth and Atmospheric Sciences at Cornell University and a co-author on the study. “This gives us another tool for trying to figure out how these oceans work. And the big question is, are things living there, or could they?”

Continue reading “Exploring Ice-Ocean Interactions on Alien Moons” »

Some of Apple’s biggest investors are set to pressure the company tomorrow to reveal its use of artificial intelligence tools (via the Financial Times).

Apple’s annual shareholder meeting takes place tomorrow, allowing those with a major stake in the company to put forward proposals. One resolution proposed by the American Federation of Labor and Congress of Industrial Organizations (AFL-CIO) asks Apple to disclose its use of AI and any ethical guidelines that the company has adopted regarding the technology.

Researchers at EPFL have developed a new, uniquely modular machine learning model for flexible decision-making. It is able to input any mode of text, video, image, sound, and time-series and then output any number, or combination, of predictions.

We’ve all heard of large language models, or LLMs—massive scale deep learning models trained on huge amounts of text that form the basis for chatbots like OpenAI’s ChatGPT. Next-generation multimodal models (MMs) can learn from inputs beyond text, including video, images, and sound.

Creating MM models at a smaller scale poses significant challenges, including the problem of being robust to non-random missing information. This is information that a model doesn’t have, often due to some biased availability in resources. It is thus critical to ensure the model does not learn the patterns of biased missingness in making its predictions.

Elon Musk claims OpenAI is using GPT-4 to ‘maximize profits’ instead of ‘for the benefit of humanity.’

The lawsuit claims that the GPT-4 model OpenAI released in March 2023 isn’t just capable of reasoning but is also actually “better at reasoning than average humans,” having scored in the 90th percentile on the Uniform Bar Examination for lawyers. The company is rumored to be developing a more advanced model, known as “Q Star,” that has a stronger claim to being true artificial general intelligence (AGI).

Altman was fired (and subsequently rehired five days later) by OpenAI in 2023 over vague claims that his communication with the board was “hindering its ability to exercise its responsibilities.” The lawsuit filed by Musk alleges that in the days following this event, Altman, Brockman, and Microsoft “exploited Microsoft’s significant leverage over OpenAI” to replace board members with handpicked alternatives that were better approved of by Microsoft.

Continue reading “Elon Musk sues OpenAI for abandoning its mission to benefit humanity” »