Lifeboat Foundation

AI systems, such as GPT-4, can now learn and use human language, but they learn from astronomical amounts of language input—much more than children receive when learning how to understand and speak a language. The best AI systems train on text with a word count in the trillions, whereas children receive just millions per year.

Due to this enormous data gap, researchers have been skeptical that recent AI advances can tell us much about human learning and development. An ideal test for demonstrating a connection would involve training an AI model, not on massive data from the web, but on only the input that a single child receives. What would the model be able to learn then?

Continue reading “New research shows how child-like language learning is possible using AI tools” »

The amount of digital data available is greater than ever before, including in health care, where doctors’ notes are routinely entered into electronic health record systems. Manually reviewing, analyzing, and sorting all these notes requires a vast amount of time and effort, which is exactly why computer scientists have developed artificial intelligence and machine learning techniques to infer medical conditions, demographic traits, and other key information from this written text.

However, safety concerns limit the deployment of such models in practice. One key challenge is that the medical notes used to train and validate these models may differ greatly across hospitals, providers, and time. As a result, models trained at one hospital may not perform reliably when they’re deployed elsewhere.

Previous seminal works by Johns Hopkins University’s Suchi Saria—an associate professor of computer science at the Whiting School of Engineering—and researchers from other top institutions recognize these “dataset shifts” as a major concern in the safety of AI deployment.

Saint Louis University associate professor of health management and policy in the College for Public Health and Social Justice, SangNam Ahn, Ph.D., recently published a paper in Journal of Clinical Psychology that examines the relationship between childhood adversity, and psychiatric decline as well as adult adversity and psychiatric and cognitive decline.

His team discovered that just one instance of adversity in childhood can increase cases of mental illness later in life, and adverse events in adults can lead to a greater chance of both mental illness and cognitive decline later in life.

“Life is very complicated, very dynamic,” Ahn said. “I really wanted to highlight the importance of looking into the lasting health effect of adversity, not only childhood but also adulthood adversity on health outcomes, especially physical health and psychiatric and cognitive health. There have been other studies before, but this is one of the first that looks into these issues comprehensively.”

Johns Hopkins Medicine neuroscientists say they have found a new function for the SYNGAP1 gene, a DNA sequence that controls memory and learning in mammals, including mice and humans.

The finding, published in Science, may affect the development of therapies designed for children with SYNGAP1 mutations, who have a range of neurodevelopmental disorders marked by intellectual disability, autistic-like behaviors, and epilepsy.

In general, SYNGAP1, as well as other genes, control learning and memory by making proteins that regulate the strength of synapses—the connections between brain cells.

Researchers at the Centre for Genomic Regulation (CRG) have found that the Snhg11 gene is critical for the function and formation of neurons in the hippocampus. Experiments with mice and human tissues revealed that the gene is less active in brains with Down syndrome, potentially contributing to the memory deficits observed in people living with the condition. The findings are published in the journal Molecular Psychiatry.

Traditionally, much of the focus in genomics has been on protein-coding genes, which in humans constitute around just 2% of the entire genome. The rest is “dark matter,” including vast stretches of non-coding DNA sequences that do not produce proteins but are increasingly recognized for their roles in regulating gene activity, influencing genetic stability, and contributing to complex traits and diseases.

Snhg11 is one gene found in the dark matter. It is a long non-coding RNA, a special type of RNA molecule that is transcribed from DNA but does not encode for a protein. Non-coding RNAs are important regulators of normal biological processes, and their abnormal expression has been previously linked to the development of human diseases, such as cancer. The study is the first evidence that a non-coding RNA plays a critical role in the pathogenesis of Down syndrome.

Artificial intelligence is pivotal in advancing biotechnology and medical procedures, ranging from cancer diagnostics to the creation of new antibiotics. However, the ecological footprint of large-scale AI systems is substantial. For instance, training extensive language models like ChatGPT-3 requires several gigawatt-hours of energy—enough to power an average nuclear power plant at full capacity for several hours.

Prof. Mario Chemnitz and Dr. Bennet Fischer from Leibniz IPHT in Jena, in collaboration with their international team, have devised an innovative method to develop potentially energy-efficient computing systems that forego the need for extensive electronic infrastructure.

They harness the unique interactions of light waves within optical fibers to forge an advanced artificial learning system. Unlike traditional systems that rely on computer chips containing thousands of electronic components, their system uses a single optical fiber.

Doroni’s H1X should help you beat traffic with a 120 mph top speed.

If you are not clued up on Arrakis it is a desert planet that is home to a valuable mineral in the Dune universe called ‘spice’ which is mined using giant harvesters. As you can imagine being a desert planet isn’t a desirable place to be and to make it worse, water is such a scarce commodity that people have to wear suits that will harness the body’s water so it can be replenished and drank.

If that wasn’t bad enough, there are also gigantic sandworms that roam the deserts of Arrakis, which to put it bluntly, you just don’t want to encounter.

Auricular acupuncture, recommended by the World Health Organization (WHO) and offered as an integrative practice since 2006 by the SUS (Sistema Única de Saúde), Brazil’s national health service, is safe for patients with depression and effectively reduces symptoms of this mental health disorder, according to a study conducted by researchers at the University of São Paulo (USP) and the University of Southern Santa Catarina (UNISUL).

The results of the study are reported in an article published in the journal JAMA Network Open. They confirm the efficacy of auricular acupuncture as an alternative treatment for depression, a mood disorder for which rising numbers are seeking care from the SUS, judging from data provided by the Ministry of Health.

Depression is one of the leading causes of disability worldwide, according to the WHO. In Brazil, the lifetime prevalence of depression is 15.5%, one of the highest globally, and depressive disorders account for 10.3% of years of life lost (YLL), a measure of premature mortality calculated by subtracting the age at death from the longest possible life expectancy for a person at that age.