(Press-News.org) When you teach a child how to solve puzzles, you can either let them figure it out through trial and error, or you can guide them with some basic rules and tips. Similarly, incorporating rules and tips into AI training—such as the laws of physics—could make them more efficient and more reflective of the real world. However, helping the AI assess the value of different rules can be a tricky task.
Researchers report March 8 in the journal Nexus that they have developed a framework for assessing the relative value of rules and data in “informed machine learning models” that incorporate both. They showed that by doing so, they could help the AI incorporate basic laws of the real world and better navigate scientific problems like solving complex mathematical problems and optimizing experimental conditions in chemistry experiments.
“Embedding human knowledge into AI models has the potential to improve their efficiency and ability to make inferences, but the question is how to balance the influence of data and knowledge,” says first author Hao Xu of Peking University. “Our framework can be employed to evaluate different knowledge and rules to enhance the predictive capability of deep learning models.”
Generative AI models like ChatGPT and Sora are purely data-driven—the models are given training data, and they teach themselves via trial and error. However, with only data to work from, these systems have no way to learn physical laws, such as gravity or fluid dynamics, and they also struggle to perform in situations that differ from their training data. An alternative approach is informed machine learning, in which researchers provide the model with some underlying rules to help guide its training process, but little is known about the relative importance of rules vs data in driving model accuracy.
“We are trying to teach AI models the laws of physics so that they can be more reflective of the real world, which would make them more useful in science and engineering,” says senior author Yuntian Chen of the Eastern Institute of Technology, Ningbo.
To improve the performance of informed machine learning, the team developed a framework to calculate the contribution of an individual rule to a given model’s predictive accuracy. The researchers also examined interactions between different rules because most informed machine learning models incorporate multiple rules, and having too many rules can cause models to collapse.
This allowed them to optimize models by tweaking the relative influence of different rules and to filter out redundant or interfering rules entirely. They also identified some rules that worked synergistically and other rules that were completely dependent on the presence of other rules.
“We found that the rules have different kinds of relationships, and we use these relationships to make model training faster and get higher accuracy,” says Chen.
The researchers say that their framework has broad practical applications in engineering, physics, and chemistry. In the paper, they demonstrated the method’s potential by using it to optimize machine learning models to solve multivariate equations and to predict the results of thin layer chromatography experiments and thereby optimize future experimental chemistry conditions.
Next, the researchers plan to develop their framework into a plugin tool that can be used by AI developers. Ultimately, they also want to train their models so that the models can extract knowledge and rules directly from data, rather than having rules selected by human researchers.
“We want to make it a closed loop by making the model into a real AI scientist,” says Chen. “We are working to develop a model that can directly extract knowledge from the data and then use this knowledge to create rules and improve itself.”
###
This research was supported by the National Center for Applied Mathematics Shenzhen, the Shenzhen Key Laboratory of Natural Gas Hydrates, the SUSTech – Qingdao New Energy Technology Research Institute, and the National Natural Science Foundation of China.
Nexus, Xu et al., “Worth of prior knowledge for enhancing deep learning” https://www.cell.com/nexus/fulltext/S2950-1601(24)00001-9
END
Planting new coral in degraded reefs can lead to rapid recovery – with restored reefs growing as fast as healthy reefs after just four years, new research shows.
Reefs worldwide are severely threatened by local and global pressures. In Indonesia, where the study was carried out, destructive blast fishing destroyed large reef areas 30-40 years ago – with no signs of recovery until now.
The Mars Coral Reef Restoration Programme attempts to restore degraded reefs by transplanting coral fragments onto a network of interconnected ...
About The Study: In this study including 47,000 participants, continuous Medicaid eligibility during the COVID-19 pandemic significantly reduced loss of Medicaid after birth, suggesting similar uninsurance reductions may be expected from post-pandemic postpartum Medicaid extensions, which most states plan to implement.
Authors: Jamie R. Daw, Ph.D., of the Columbia University Mailman School of Public Health in New York, is the corresponding author.
To access the embargoed study: Visit our For The Media website at this link https://media.jamanetwork.com/
(doi:10.1001/jamahealthforum.2024.0004)
Editor’s ...
About The Study: This qualitative study including 25 interviews with Black patients discharged from the emergency department described these patients’ perspectives about racism in health care, recent clinical experiences, and thoughts on system improvements. Black patients described a notable amount of medical mistrust, anticipation of racism in emergency care, and personal experiences with clinical instances of racism in emergency treatment.
Authors: Anish K. Agarwal, M.D., M.P.H., M.S., of the University of Pennsylvania in Philadelphia, is the corresponding ...
Background and Aims
The impact of the characteristics of extrahepatic organ failure (EHOF) including the onset time, number, type, and sequence on the prognosis of acute-on-chronic liver failure (ACLF) patients remains unknown. This study aimed to identify the association between the characteristics of EHOF and the prognosis of ACLF patients.
Methods
ACLF subjects enrolled at six hospitals in China were included in the analysis. The risk of mortality based on the characteristics of EHOF was evaluated. Survival of study groups was ...
UVA Health researchers have received $2.8 million to advance their development of a high-tech way to save heart and dialysis patients from the need for multiple surgeries.
Researchers Lian-Wang Guo, PhD, and K. Craig Kent, MD, are pioneering a quick and gentle technique to “paint” tiny nanoparticles on transplanted veins to prevent the veins from becoming clogged in the future. This type of blockage often causes cardiovascular and dialysis patients to require repeated surgeries; approximately half of all heart ...
The human vascular system, a complex network of blood vessels, plays an essential role in maintaining health. Understanding the molecular mechanisms underlying vascular development is important for tackling age-related disorders. The Forkhead box O (FOXO1) transcription factor is crucial in processes related to aging, cellular metabolism, and apoptosis. Despite its significance, the comprehensive regulation of FOXO1 across the genome in endothelial cells (EC) has not been investigated.
In a new paper published in iScience in February 2024, a research team from Kumamoto University ...
Scope
For decades, robotic systems have played a pivotal role in nurturing the growth of tissue-engineered constructs in controlled environments through the provision of mechanical stimulation. The importance of physical stresses in tissue maturation is not only intuitive from our own bodily experiences but is also supported by a growing body of mechanotransduction research. However, experimental studies have predominantly remained confined to basic in vitro setups, hampering our ability to produce functional grafts that can translate into clinical practice. Emerging evidence suggests that replicating physiological stresses more faithfully could further enhance the functionality of tissue ...
mRNA vaccines developed against the spike glycoprotein of severe acute respiratory syndrome type 2 coronavirus (SARS-CoV-2), displayed remarkable efficiency in combating coronavirus 19 (COVID-19). These vaccines work by triggering both cellular and humoral immune responses against the spike protein of the virus. Cellular immunity may play a more protective role than humoral immunity to variants of concerns (VOC) against SARS-CoV-2, as it targets the conserved regions of spike protein and possibly cross-reacts with other variants.
Since a single spike epitope is recognized by multiple T-cell clones, the ...
Climate change is driving both the loss of biodiversity and the need for clean, renewable energy. It is also shifting where species are expected to live in the future. Yet these realities are rarely considered together. Where can clean energy projects be built without impacting the future habitat ranges of threatened and endangered species?
A study from the University of California, Davis, examines this question by overlaying renewable energy siting maps with the ranges of two species in the southwestern United States: the iconic and climate-vulnerable ...
In the space of just a few seconds, a person walking down a city block might check their phone, yawn, worry about making rent, and adjust their path to avoid a puddle. The smell from a food cart could suddenly conjure a memory from childhood, or they could notice a rat eating a slice of pizza and store the image as a new memory.
For most people, shifting through behaviors quickly and seamlessly is a mundane part of everyday life.
For neuroscientists, it’s one of the brain’s most remarkable capabilities. That’s because different activities require the brain to use different combinations of its many regions and billions of neurons. How ...
