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 41–60
Recent research published on arXiv today delves into the fundamental mechanisms behind large language model (LLM) hallucinations, the complexities of multi-stakeholder alignment, and the critical gap in formal specification autoformalization for AI coding agents. These three pape...
Read article →The innovation engine continues to hum, with deep tech and logistics platforms recently securing substantial investment rounds, signaling robust investor confidence in foundational technologies. OpenRouter, an emerging player in the multi-AI-model ecosystem, has more than doubled...
Read article →A new paper published on arXiv challenges a foundational assumption in AI for Science (AI4Science), arguing that the complex, multi-stage processes generating scientific datasets should be treated as integral inference components, rather than fixed, unalterable inputs arXiv CS. L...
Read article →New research from arXiv CS. LG delivers a crucial insight into the cognitive architecture of Large Language Models (LLMs), revealing they currently exhibit 'no signs of individuated metacognition' – a critical ability to assess one's own capabilities without explicitly exercising...
Read article →A significant collection of machine learning research, made public today on arXiv, reveals a broad spectrum of advancements poised to enhance the reliability, interpretability, and efficiency of AI systems. These papers tackle fundamental challenges, from making algorithms more r...
Read article →A fresh wave of theoretical machine learning research, published today on arXiv CS. LG, signals a focused effort by the AI community to deepen our understanding of neural network behavior....
Read article →A new research paper, "MEDAL: Manifold Embedding Distillation via Autoencoder Learning," has been published on arXiv, addressing critical limitations in popular nonlinear dimension reduction methods like t-SNE and UMAP. The paper introduces a novel approach aimed at providing rig...
Read article →A new wave of AI research, unveiled today on arXiv, is fundamentally reshaping how intelligent systems analyze and represent complex data. Leading this charge are innovations such as a faithful knowledge base embedding method, BoxLitE, and a noise-robust system for understanding ...
Read article →A wave of new research papers published on arXiv today reveals significant progress in overcoming persistent challenges across computer vision and robotics, tackling everything from the fundamental 'binding problem' in visual understanding to energy-efficient perception for auton...
Read article →Recent research released on arXiv highlights significant strides in making large language models (LLMs) and their agents more robust, safe, and efficient for real-world applications. These advancements tackle critical challenges like mitigating hallucinations, ensuring fairness a...
Read article →A new research paper published on arXiv introduces a novel approach to AI agent decision-making, focusing on a 'free exploration budget' to minimize regret in complex environments. This paradigm shift, outlined in arXiv:2605....
Read article →A trio of new research papers, all published today on arXiv, signal significant advancements in addressing some of the most persistent challenges facing Graph Neural Networks (GNNs). These breakthroughs tackle GNN robustness against adversarial attacks, enhance their ability to c...
Read article →A flurry of recent research, published today on arXiv CS. LG, reveals groundbreaking insights into the fundamental scaling laws of neural networks, particularly focusing on the crucial role of sparse activations....
Read article →SpaceX's ambitious Starship program has seen its latest iteration, Starship V3, complete a "mostly successful" inaugural flight, a crucial step forward for the heavy-lift vehicle, though engineers acknowledge significant work remains before it achieves low-Earth orbit capabilitie...
Read article →A flurry of new research papers, primarily from arXiv CS. AI and released on May 23, 2026, signals a concerted effort to push AI beyond static data and into the nuanced complexities of the physical world....
Read article →Deep learning, traditionally rooted in numerical optimization, is seeing a fascinating wave of foundational research exploring alternative paradigms and enhancing existing architectures. Recent papers, all announced on arXiv on May 23, 2026, reveal explorations into algebraic lea...
Read article →A torrent of new research, primarily from arXiv's latest releases on May 23, 2026, reveals a profound acceleration in artificial intelligence's capacity for sophisticated data analysis and synthesis. This wave of breakthroughs is not merely about pushing performance metrics, but ...
Read article →Today's flurry of new research papers unveils a fascinating dual trajectory in the evolution of large language models (LLMs) and multimodal large language models (MLLMs): on one hand, we're seeing impressive leaps in their reasoning capabilities and real-world applicability; on t...
Read article →A significant wave of new research papers, all published today, reveals a burgeoning paradigm in scientific discovery: autonomous, multi-agent AI systems are rapidly evolving from assistive tools to genuine “co-scientists. ” These advanced AI agents are demonstrating capabilities...
Read article →Today, three significant research papers published on arXiv CS. AI collectively point towards a future of more efficient, scalable, and robust AI agents....
Read article →