Academic papers, research breakthroughs, and scientific advances in machine learning and AI.
319 articles
Researchers from Columbia University and the University of Cambridge have created a new benchmark to test how well machine learning models predict material properties. The work aims to improve the reliability of AI-driven materials science by combining quantum computing insights with ML evaluation methods. Source: quantumzeitgeist.com
The credit score, a single three-digit number, heavily influences access to mortgages, car loans, and other credit in the US. Machine learning is increasingly being applied to refine how these scores are calculated and used by lenders. Source: cacm.acm.org
Quantum machine learning combines quantum computing with AI techniques, promising faster processing for certain tasks through hybrid quantum-classical models. The field is still young, with most current implementations limited by hardware constraints and far from replacing traditional ML approaches. Source: usa.inquirer.net
Google DeepMind has released the AlphaGenome Atlas, a free research resource offering precomputed molecular effect predictions and AVI scores for roughly 9 billion human DNA variants. The tool aims to help scientists quickly assess how genetic variations may influence biological function. Source: marktechpost.com
Researchers at Princeton have built an AI called PACMAN that can steer fusion plasma in real time, reacting within milliseconds. The system is also able to forecast dangerous plasma instabilities before they occur, potentially improving fusion reactor safety. Source: scitechdaily.com
A research group led by Michele Simoncelli has developed a new benchmark to test how accurately machine learning models capture quantum-level atomic interactions. The goal is to push AI systems used in materials science toward more physically consistent predictions. Source: eurekalert.org
AI research agents are now capable of proposing, running, and scoring their own machine learning experiments, but generating ideas is far cheaper than verifying them. Meta FAIR's new RPMs aim to prioritize promising experiments before committing costly compute resources. Source: marktechpost.com
The piece recalls how, in mid-2023, the internal “RLSlow” research project produced the first results suggesting that scaling up reasoning-focused training was actually feasible. That early milestone reportedly marked a turning point in the team's confidence about pushing reasoning capabilities further. Source: openai.com
Researchers from UCLA and the University of Rochester have advanced an imaging system that uses AI tailored to physical signal processing rather than typical digital data. The approach could enhance detail capture in biomedical imaging as well as sensors used in autonomous vehicles. Source: technology.org
A new study explores how machine learning can improve the way mobile networks manage radio resources, a critical and limited asset in 5G systems. The proposed framework focuses on smarter, more adaptive allocation methods to boost network efficiency. Source: bioengineer.org
A new study used a neural network model to examine how different types of training influence the learning process itself. The findings shed light on mechanisms of learning that may apply both to artificial and biological systems. Source: healthcare.utah.edu
Google Research engineers Joey Poomarin Phloyphisut and Cory McLean describe a transfer learning approach aimed at improving genomic prediction accuracy for populations historically underrepresented in genetic studies. The work seeks to reduce bias in genomic models trained mostly on limited demographic data. Source: research.google
The National Science Foundation is backing a new free, open-source software project that uses machine learning to help materials scientists answer a common research question faster. The tool aims to make discovery workflows more accessible to the wider scientific community. Source: news.engineering.tamu.edu
A new study explores machine learning models to predict retinopathy of prematurity, an eye condition that can cause vision loss in newborns. Early detection through such models could help clinicians intervene sooner to prevent blindness. Source: nature.com
Researchers are using machine learning to build continuously updated 'living' maps of evidence, aiming to address persistent health inequalities in primary care. The approach could help identify where research and treatment gaps hit disadvantaged patients hardest. Source: bioengineer.org
New research challenges the idea that bigger models and more data automatically mean better financial predictions. Instead, combining traditional econometric methods with machine learning — while carefully controlling model capacity — improves volatility forecasting accuracy. Source: bioengineer.org
Researchers at the University of Miami developed machine-learning models that combine clinical, biomarker, and imaging data to identify Parkinson's patients likely to decline more rapidly. The approach could help doctors intervene earlier with more targeted treatment. Source: newswise.com
Google's new TimesFM-3 is a foundation model built for time series forecasting that can handle multiple variables at once without task-specific training. It aims to deliver state-of-the-art accuracy across diverse forecasting scenarios in a single unified approach. Source: research.google
A machine-learning tool called symclatron analyzes genomic sequencing data to predict whether uncultivated bacteria and archaea live independently or form symbiotic relationships with other organisms. The framework offers a new way to study microbes that can't be grown in lab conditions. Source: nature.com
Researchers built a behavior-based AI system that analyzes USDT and USDC transactions on Ethereum, classifying wallets as sanctioned/frozen, linked to cybercrime, or legitimate. The tool aims to help authorities trace illicit stablecoin flows more effectively. Source: eurekalert.org