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 681–700
The world of artificial intelligence is abuzz with activity, and recent preprints emerging from arXiv point to a fascinating new phase in the development of Unified Multimodal Models (UMMs) and Multimodal Large Language Models (MLLMs). These papers, all published on March 23, 202...
Read article →Two new research papers, published on arXiv just today, March 23, 2026, illuminate crucial advancements in how we interact with and optimize large language models (LLMs) and integrate multimodal sensor data. These breakthroughs tackle fundamental challenges in AI deployment, from...
Read article →Researchers have unveiled a critical limitation in how we currently assess bias in face recognition systems, revealing that existing evaluation methods often miss complex, intersectional subpopulations. This groundbreaking work, detailed in a new paper from arXiv CS....
Read article →A trio of research papers, all published on arXiv on March 23, 2026, collectively point to a significant new direction in artificial intelligence: the pervasive use of dynamic, graph-based knowledge representation to enable more adaptive Large Language Models (LLMs) and a deeper ...
Read article →New research published on arXiv unveils a suite of AI breakthroughs poised to transform diverse areas of healthcare, from predicting dementia progression and diagnosing complex brain disorders to modeling intricate RNA structures and automating cardiac diagnostics. These papers, ...
Read article →The world of AI research is buzzing with a fresh wave of foundational breakthroughs, as evidenced by a flurry of papers published on arXiv on March 23, 2026. These developments collectively point towards a future where AI models are not only more capable but also significantly mo...
Read article →Recent research published on arXiv CS. AI on March 23, 2026, unveils significant advancements in applying artificial intelligence, particularly deep reinforcement learning (DRL), to solve some of the most challenging problems in resource management and operational logistics....
Read article →Breakthroughs in Geometric Deep Learning (GDL) are extending AI’s reach beyond traditional grid-like data, with recent research highlighting new theoretical frameworks and concrete applications in fields like drug discovery. A flurry of new papers published on arXiv this week und...
Read article →A flurry of groundbreaking research papers, all announced on March 23, 2026, signals a vibrant and multi-faceted evolution in Large Language Model (LLM) development. These studies, spanning from fundamental improvements in how LLMs process language to novel applications in hardwa...
Read article →A new wave of research is pushing the boundaries of artificial intelligence, with Geometric Deep Learning (GDL) and Graph Neural Networks (GNNs) rapidly evolving to tackle complex, non-Euclidean data structures. Recent arXiv publications highlight significant progress, particular...
Read article →A flurry of new research papers on arXiv, all announced or updated on March 23, 2026, highlights significant strides in making Large Language Models (LLMs) more reliable, efficient, and capable across diverse applications. From novel methods to combat hallucinations to breakthrou...
Read article →The frontier of artificial intelligence is expanding at an exhilarating pace, marked by the emergence of sophisticated new architectures designed for complex automation. Today, arXiv introduces "Autonoma," a novel hierarchical multi-agent framework that promises to translate open...
Read article →The sheer velocity of AI research is undeniable, with the International Joint Conference on Neural Networks (IJCNN) 2025 reporting a 100% growth in paper submissions and a 200% increase in active reviewers arXiv CS. LG....
Read article →A wave of new research papers published today on arXiv highlights critical strides in making large language models safer, more efficient, and adaptable to complex real-world challenges. From tackling unpredictable side effects in model editing to optimizing visual processing in m...
Read article →A groundbreaking discovery published on arXiv today suggests that the key-value (KV) cache, long considered an essential component for efficient transformer inference in Large Language Models (LLMs), is entirely redundant. Researchers demonstrate that keys and values at every lay...
Read article →A wave of new foundational research, announced today, significantly advances AI's ability to operate effectively and efficiently in complex, real-world environments by tackling challenges like uncertainty, spurious correlations, and computational overhead. These advancements, doc...
Read article →The landscape of scientific discovery is shifting, with new research unveiled on arXiv this week demonstrating a profound evolution in how AI interacts with and contributes to the scientific process. Far from merely predicting outcomes, AI models are now advancing capabilities in...
Read article →A trio of significant research preprints, all published on arXiv on March 23, 2026, collectively illuminate the cutting edge of AI development, spanning robotics in challenging environments, scalable speech generation, and critical insights into medical AI foundation models. Thes...
Read article →A comprehensive wave of foundational artificial intelligence and machine learning research has emerged from arXiv (Computer Science), with 48 distinct papers published simultaneously on 2026-03-05. This synchronous release signifies an accelerated expansion of the theoretical and...
Read article →A significant collection of eight research papers, released concurrently on 2026-03-06, outlines substantial advancements in the fundamental methodologies of Artificial Intelligence model training and optimization. These developments collectively enhance efficiency, stability, an...
Read article →