Articles tagged with Artificial Intelligence
A study by Georgia State University found that people rate AI responses as more virtuous and intelligent than human responses on moral issues. The AI system, ChatGPT, was able to fool participants into believing its responses were better, highlighting the need for a deeper understanding of AI's role in society.
Recent technological breakthroughs enable global insect surveys using sound recordings, images, radar, laser beams, and DNA analysis. This allows for new insights into insect trends and species identification without relying on experts or killing insects.
A new study published in JMIR Mental Health sheds light on the promising role of virtual reality (VR) in treating major depressive disorder (MDD). VR-Enhanced Behavioral Activation (XR-BA) was found to be effective in easing symptoms of depression, comparable to traditional behavioral activation.
Researchers created GraSSRep and rhea, tools that outperform current methods for handling repeats and structural variants in metagenomic data. These methods use self-supervised learning and graph neural networks to analyze microbiome data, offering new insights into biological processes and potential applications in antibiotic resistance.
Researchers caution that current AI-based deception detection methods are plagued by a lack of explainability, risk of biased results, and deficits in theoretical foundation. They emphasize the need for controlled laboratory experiments, large and diverse data sets, and validation on independent datasets to ensure accurate results.
A new AI method developed by Swedish researchers can identify toxic substances based on their chemical structure, potentially replacing animal testing. The method has been shown to be more accurate and broadly applicable than existing computational tools, offering a promising alternative for environmental research and authorities.
Research suggests that four proposed AI search engines for automating search and retrieval of digital histopathology slides have inadequate performance for routine clinical care. The algorithms showed less than 50% accuracy in some cases, highlighting the need for rigorous external validations and standardization before clinical adoption.
A new AI model developed by researchers at the University of British Columbia can accurately predict if a patient receiving cancer care will require mental health services. The AI analyzes oncologist's notes and identifies subtle clues that suggest a patient may benefit from early psychiatric or counselling interventions.
Researchers at Universitat Oberta de Catalunya have developed a new AI system that can detect asbestos in roof images with high accuracy, enabling its removal. The system uses freely available aerial photographs and has demonstrated an accuracy of over 80%, making it more scalable and adaptable than previous methods.
An interdisciplinary team has published guidelines for responsible use of machine learning in science, focusing on transparency and reproducibility. The checklist includes detailed descriptions of machine learning models, data, and experimental designs to ensure integrity and validity.
FathomVerse combines real underwater images, interactive gameplay, and AI training to improve ocean research. The game engages players in labeling images for machine learning models, contributing to a broader audience in ocean exploration.
Researchers found that chatbots like ChatGPT can provide accurate information on climate change-related topics, but not all responses are reliable. The study compared the chatbots' answers to hazard risk indices generated by the Intergovernmental Panel on Climate Change and found more accuracy with tropical storms than droughts.
A new study found that GPT-4 and Google Gemini performed poorly in breast imaging classification, with a high percentage of discordant assignments that could impact patient management. The researchers emphasize the need for regulation of these large language models in high-level medical scenarios.
The US Preventive Services Task Force recommends biennial breast cancer screenings starting at age 40 for all women, aiming to enhance early detection and tackle disparities. However, challenges remain in ensuring equitable access to screening technologies and addressing gaps in evidence regarding AI-powered image interpretation.
Scientists from Penn created a non-volatile memory device using ferroelectric aluminum scandium nitride (AlScN) to retain data at high temperatures. The device's stability and fast switching properties enable efficient computation in harsh conditions, including space exploration and deep-earth drilling.
A new study found that ChatGPT's responses to key questions about brain abscess diagnostics and treatment were unclear or incorrect in some cases, potentially putting patients at risk. The AI model provided clear answers to most questions but failed to provide adequate guidance on certain treatment decisions.
A new branch of AI, causal machine learning (ML), can estimate treatment outcomes better than traditional methods. This enables personalized treatment strategies and improved patient health.
Researchers from NIMS and Tokyo University of Science developed a compact AI device that utilizes molecular vibrations to predict blood glucose levels in patients with diabetes. The device outperformed existing AI devices with a 50% error reduction, paving the way for low-power AI terminal devices with various sensors.
Researchers at Princeton University have developed a device that improves image quality in holographic displays, enabling a wider field of view and more stable images. The new technology has the potential to transform various industries, from healthcare to home repairs.
A study by Mass General Brigham demonstrates that large language models can help reduce physician workload and improve patient education, but also identifies limitations that may affect patient safety. The researchers emphasize the need for vigilant oversight and a measured approach to LLM implementation.
Using AI software called SLEAP, scientists at Salk Institute are designing climate-saving plants with optimized root systems that can store more carbon. This approach enables researchers to analyze plant features and connect desirable traits to targetable genes, accelerating the development of carbon-capturing plants.
Research at Tampere University aims to develop adaptable safety systems for autonomous off-road mobile machinery. The framework focuses on ensuring public safety without compromising technological advancements, incorporating a human-in-the-loop safety option.
The ECOG-ACRIN Cancer Research Group will present late-breaking findings on prostate cancer, breast cancer, and pancreas cancer at ASCO 2024. These studies include optimal chemotherapy for Black women with breast cancer and results from the first randomized national trial in older adults with pancreas cancer.
A new study from Binghamton University explores how artificial intelligence can improve flexibility and efficiency in customer service centers. The research found that AI systems can help organizations balance exploiting the latest technology to boost efficiency with exploring new areas to gain insights.
A Cleveland Clinic study uses reinforcement learning to design antibiotic regimens that minimize resistance and maximize susceptibility. The AI model predicts the most efficient treatment plans for multiple strains of E. coli, informing hospital-wide infection management.
Researchers from Drexel University have developed a machine-learning algorithm that can extract and recognize digital fingerprints of AI-generated video, including Stable Video Diffusion, Video-Crafter, and Cog-Video. The algorithm can be trained on just a few examples of new AI generators to detect them.
The position paper identifies eight key areas for digital neuroscience research, including near-term, middle-term, and long-term goals. It also discusses the potential of 'digital twin' approaches, ultra-high-resolution digital atlases, and neuro-derived AI and computing innovations.
A new AI model generates realistic images of single cells, which are used as synthetic data to train an AI model for better cell segmentation. The researchers found that providing a more diverse dataset during training improves performance.
Researchers used machine learning to analyze over 1.09 million potential metabolite-receptor pairs and predict the likelihood that each interaction contributed to Alzheimer's disease. The study found a protective metabolite called agmatine interacting with a receptor called CA3R in brain cells, reducing CA3R levels and lowering phospho...
Researchers used freely available AI chatbots to simulate a multidisciplinary workshop and found consistency in responses suggesting an online platform for pain science education. The study highlights the potential of AI to contribute to educational strategies and clinical research, leveling the field for resource-limited settings.
A new study by Mass General Brigham researchers found that larger 'foundation models' can help mitigate disparities between different demographic groups and enhance model accuracy. The findings emphasize the need for more diverse training datasets and demographic-stratified evaluations of AI systems.
Scientists have found evidence of a Bragg glass phase in a crystal using machine learning and X-ray technology. The study provides insight into the nature of glasses and their unique properties, which could inform material design.
Researchers at MIT designed a communication system called LiTweeture that allows users to selectively reveal verified aspects of their identity, reducing social inhibition and encouraging junior academics to participate in public discourse. This approach has the potential to be applied to any community or discussion space.
A study led by the University of Cambridge found that AI models like GPT-4 outperformed doctors in diagnosing and treating eye problems. The researchers suggest using AI for triaging patients and providing advice in well-controlled contexts, where access to specialist healthcare professionals is limited.
A new study found that ChatGPT can consistently recommend deprescribing medications in patients without a history of cardiovascular disease. However, the AI model was more cautious when considering patients with cardiovascular disease and varied its responses depending on the chat session.
DZNE researchers aim to measure intricate brain structures with artificial intelligence, using 3D brain scans and data from the Human Connectome Project. The goal is to better understand disease processes and normal brain changes over time.
Researchers from the University of Cambridge used AI to identify compounds that block alpha-synuclein aggregation, a key step in treating Parkinson's disease. This breakthrough could lead to faster development of new treatments for the condition, which affects over six million people worldwide.
A new AI-based model at Stanford Hospital predicts patient decline and flags physicians and nurses, enhancing communication and intervention. The study saw a significant improvement in clinical outcomes, with a 10.4% decrease in deterioration events among high-risk patients.
A study found that GPT-4 matched the performance of radiologists in detecting errors in radiology reports, with an accuracy rate of 82.7%. The use of GPT-4 resulted in lower mean correction cost per report than the most cost-efficient radiologist.
The authors suggest a regulatory approach that accommodates the unique characteristics of generalist AI models while ensuring patient safety and supporting physician decision-making. This is crucial as rigid frameworks may hinder progress in AI-driven healthcare, while hybrid approaches could be viable solutions.
A new international research project, RoboSapiens, aims to make industrial robots more adaptable and reliable by utilizing artificial intelligence. The project, led by Aarhus University, will focus on four use-cases: autonomous ships, human-robot interactions, warehouse robots, and laptop refurbishment.
Researchers at UPV and iPRONICS have developed a groundbreaking photonic chip that can implement twelve basic functionalities required by various systems. This chip allows on-demand programming and interconnecting wireless and photonic segments of communication networks, increasing efficiency and reducing bottlenecks.
A new study reveals that generative AI can draft empathetic responses to patient messages, reducing burnout among physicians. The tool provides a starting point for physicians to craft thoughtful responses, improving communication quality and efficiency.
A new study by RMIT University found that only over a third of media organizations have an image-specific AI policy in place, highlighting the need for clearer guidelines. The research highlights the challenges of navigating generative AI technologies in visual journalism, including algorithmic bias and copyright concerns.
Researchers at PolyU are using GeoAI to monitor and analyze environmental changes, detect disaster-damaged buildings, and optimize traffic flow. The technology has the potential to support sustainable urban development and improve public health.
Mayo Clinic researchers have invented a new class of artificial intelligence (AI) algorithms called hypothesis-driven AI, which can help discover the complex causes of diseases like cancer and improve treatment strategies. This emerging class of AI offers an innovative way to use massive datasets to guide individualized medicine.
Researchers used an AI-assisted application to help people write cartoon captions for The New Yorker Cartoon Caption Contest. The tool analyzed incongruity and generated suggestions, resulting in jokes rated 30% funnier than those written without assistance.
The DISCount framework combines AI-powered image analysis with human analysis to quickly deliver reliable estimates of building damage and bird flock size. It has been recognized by the Association for the Advancement of Artificial Intelligence for its social impact, winning an award for best paper on AI for social impact.
Lehigh University researcher A. Emrah Bayrak explores best practices for human-AI collaboration in complex design tasks, aiming to maximize productivity and job satisfaction. His project uses models that predict human decision-making and combines it with AI's training data analysis to determine strategies for division of labor.
A new study published in PNAS found that AI-generated messages made recipients feel more 'heard' than messages generated by untrained humans, and that AI was better at detecting emotions. However, recipients reported feeling less heard when they learned a message came from AI.
A new AI model can accurately screen for post-traumatic stress disorder (PTSD) in postpartum women, offering a potential solution for early intervention and preventing progression of the disorder. The study collected narrative accounts from 1,295 women and showed the model's superiority in identifying PTSD cases.
The project, AI STORIES, aims to understand how artificial intelligence creates new narratives and challenges dominant American narratives. The research will test the hypothesis that AI-generated responses mimic sci-fi stories where AI gains human emotions.
Researchers at Insilico Medicine have developed a novel PTPN2/N1 inhibitor with improved oral absorption and robust antitumor efficacy using the company's generative AI engine. The new compound demonstrates enhanced biological activities compared to existing inhibitors, offering new treatment possibilities for cancer patients.
Two new Review articles explore AI's application in early cancer detection, highlighting its potential to enhance diagnostic accuracy and improve treatment selection. The articles emphasize the need for a robust, multi-disciplinary approach to integrate AI into medicine.
Researchers developed an AI algorithm to identify normal mammograms and ran a simulation on patient data, revealing that AI can reduce unnecessary testing while maintaining cancer detection rates. The study suggests that AI can help doctors focus on more questionable scans, reducing false positives and improving workflows.
Zero trust architecture (ZTA) aims to prevent internal attacks by restricting behavior based on resource-based security policies. AI-powered automation and orchestration can help overcome implementation challenges, relieving security personnel from manual tasks.
A new generative AI tool models the infant microbiome, predicting neurodevelopmental deficits. The Q-net model predicts which babies are at risk for cognitive deficits with high accuracy, identifying potential treatment targets.
A new machine-learning approach outperforms human testers in generating diverse prompts that trigger a wider range of undesirable responses from chatbots. The technique provides a faster and more effective way to ensure the safety of large language models.
A pilot study found that ChatGPT provided factual responses to common questions about vaccination and STIs, demonstrating its potential as a tool to reduce vaccine hesitancy and provide helpful advice on sexual health. The AI chatbot's accuracy and conciseness were comparable to professional organizations' guidelines.
A new study by Carey Morewedge and colleagues found that people are more likely to recognize bias in algorithmic decisions than their own. This is because algorithms can codify and amplify human bias, but also reveal structural biases in society. The research suggests ways to increase awareness of biases and correct them.