Research Correspondent
Elena Voss covers model development and the ideas behind new research. Her beat follows the distance between a promising paper and a result that holds up outside the lab. She favors clear explanations, original sources and questions that a benchmark score alone cannot answer.
2,898 published articles · Showing 601–620
Today's research landscape is buzzing with foundational insights, as three distinct yet equally vital papers emerged on arXiv, each pushing the boundaries of AI and machine learning. From securing the expansive 'Foundation Model Era' to guaranteeing prediction reliability and rev...
Read article →New research published on arXiv today reveals artificial intelligence is driving a two-pronged transformation, simultaneously prompting a fundamental re-evaluation of mathematical practice and advancing the precision of creative AI systems. These insights, released on March 27, 2...
Read article →Just this week, a flurry of research emerging from arXiv points to a profound shift in how artificial intelligence, particularly large language models (LLMs) and causal AI, is being deployed in scientific and engineering domains. No longer merely serving as intelligent assistants...
Read article →Two recent pre-print papers released on arXiv signal a significant leap forward in equipping autonomous systems with more nuanced navigation and interaction capabilities. These advancements promise to enhance safety and efficiency for everything from self-driving cars to domestic...
Read article →A significant leap in AI’s role within scientific research has been reported with the verified synthesis of self-evolving agents, capable of tackling complex tasks like program repair and scientific discovery. While demonstrating immense potential for accelerating breakthroughs,...
Read article →A flurry of groundbreaking research, published today on arXiv CS. AI, is rapidly advancing our understanding of a critical challenge in large language models (LLMs): the "confidence-faithfulness gap....
Read article →The burgeoning frontier of AI research is seeing a critical shift, moving beyond raw performance metrics to prioritize trustworthiness, reliability, and security, especially in high-stakes domains. Recent arXiv preprints, all published on March 27, 2026, illuminate significant st...
Read article →A wave of recent research papers from arXiv signals a critical shift in AI development: a concerted effort to move beyond impressive demos toward building genuinely robust, safe, and auditable AI systems, particularly autonomous agents. Published just yesterday, these papers unde...
Read article →Apple finds itself at a critical juncture, simultaneously pushing the boundaries of artificial intelligence in new consumer features while grappling with persistent security vulnerabilities affecting millions of devices. Recent reports highlight a mixed landscape: on one hand, no...
Read article →The autonomous vehicle landscape is undergoing a profound divergence, with long-term projects reaching critical junctures. Just as Sony's six-year Afeela electric vehicle rollout concludes, reportedly ending "with a thud" Wired, Tesla's purpose-built, two-seater Cybercab robotaxi...
Read article →Today, two significant research papers published on arXiv highlight crucial advancements in making AI models more reliable and their performance more accurately measurable. These breakthroughs, covering neural network robustness certification and the rigorous evaluation of large ...
Read article →A significant stride in artificial intelligence research has been unveiled with the introduction of NaviMaster, the first unified agent capable of bridging the long-standing divide between graphical user interface (GUI) and embodied navigation tasks arXiv CS. LG....
Read article →A series of recent preprints on arXiv, all published today, March 26, 2026, highlight the escalating influence of artificial intelligence across both fundamental scientific research and the practical engineering of critical infrastructure. These papers reveal not just incremental...
Read article →A new wave of research, published today on arXiv, showcases significant advancements in applying artificial intelligence and quantum computing to fundamental scientific challenges, from accelerating drug discovery to revolutionizing materials science. These papers collectively hi...
Read article →Recent pre-print publications on arXiv CS. LG, all announced on March 26, 2026, collectively demonstrate AI's expanding theoretical prowess and its immediate, practical impact across critical engineering domains....
Read article →A flurry of new research papers published today on arXiv CS. LG reveals significant strides across the spectrum of Reinforcement Learning (RL), tackling critical barriers from computational efficiency on edge devices to the complexities of real-world robotic manipulation and the ...
Read article →The bleeding edge of deep learning research is witnessing a concerted shift, with a fresh batch of arXiv pre-prints, all dated March 26, 2026, pointing towards a new era of robustness, efficiency, and interpretability in neural networks. This diverse collection, spanning foundati...
Read article →The persistent challenge of making AI models "forget" specific data points, a capability increasingly vital for data protection and ethical AI, just received a significant boost from new research. A paper published on arXiv, titled “SPARE: Self-distillation for PARameter-Efficien...
Read article →A new wave of research, published this week on arXiv, signals a pivotal shift in how the AI community evaluates and understands its most advanced models. Rather than relying solely on prediction accuracy, a growing consensus emphasizes the critical need for interpretable mechanis...
Read article →New research from arXiv CS. LG reveals a significant 'alignment tax' on large language models (LLMs), where current alignment strategies inadvertently foster 'response homogenization' and obscure crucial expressions of uncertainty, posing challenges for reliable AI deployment arX...
Read article →