1496 stories tagged AI · page 4 of 63
A new study from ArXiv cs.AI tested nine open-weight language models (135M to 3B parameters) on a 1,085-question, 16-topic benchmark designed for local deployment. Results show even the smallest models perform well on structured tasks, advancing AI democratization by enabling local, private, and cost-effective AI without cloud dependency.
A comprehensive ArXiv review systematically categorizes Graph Neural Network (GNN) techniques for link prediction, covering underlying architectures and diverse graph structures. The taxonomy aims to improve recommendation systems, fraud detection, and social network analysis by better inferring missing or future connections in data.
A new study on arXiv introduces a membership inference attack (MIA) for fine-tuned discrete diffusion language models (dLLMs), revealing that attackers can efficiently determine if specific data was used in fine-tuning. This poses significant privacy risks for users of AI systems built on these models.
A new audit framework reveals that human trainers' emotional states—such as stress or distress—can systematically bias the preference labels used in Reinforcement Learning from Human Feedback (RLHF), leading to unintended shifts in AI behavior. The study proposes methods to detect and correct this 'rater state bias.'
Researchers introduced JOR-Bench, a collection of five Japanese-language benchmarks with 1,319 problems to evaluate how well large language models (LLMs) can formulate and solve operations research (OR) problems, covering linear programming, mixed-integer programming, non-linear programming, and combinatorial optimization.
Researchers at MIT and Stanford have created novel semiconductor materials that enable AI chips to process information faster while using less energy and generating less heat. This breakthrough could make high-performance AI affordable and accessible in everyday devices like smartphones and smart home gadgets.
Google has launched Gemini 3.5 Flash Cyber, a cost-effective AI model designed to identify and fix security vulnerabilities quickly. This new tool aims to provide a more affordable alternative to larger, more expensive AI systems like Anthropic's Mythos.
A federal judge approved Anthropic's $1.5 billion settlement in a landmark copyright lawsuit over AI training data. The deal resolves one case but leaves the broader legal and ethical questions about using copyrighted works to train AI models unresolved.
A new study on the open-weight LLM Qwen3-8B reveals that AI models frequently commit to an answer before reasoning through a problem — even when the answer contradicts the task premise. Using a simple car wash scenario, researchers found the model recommended walking instead of driving in 85–100% of cases, despite driving being the only logical choice.
Researchers introduced Generative Ontology Induction (GOI), a domain-agnostic framework that automatically creates structured knowledge systems from text corpora. GOI identifies entities, dimensions, properties, relationships, and constraints, then exports them as a typed graph in YAML/JSON — making AI systems more adaptable to new topics without manual programming.
Researchers introduced SkillCorpus, a framework to consolidate and evaluate the fragmented open-source SKILL.md ecosystem for LLM agents, aiming to improve real-world agent task performance.
Provena is an open-source library that governs the data flowing into AI agents' context windows—filling a critical gap in AI governance. The project has 7 contributors and is actively seeking more, with 11 'good first issue' labels and 17 'help wanted' issues.
Nvidia is building Japan's first national AI supercomputer, called FRONTIER, to accelerate scientific research in climate science, medicine, and materials science. The project marks a major step in national AI infrastructure and raises questions about how countries should invest in sovereign AI capabilities.
Researchers have introduced Causal-Audit, a framework that enables AI models to explicitly show their step-by-step reasoning about cause-and-effect relationships, making AI decisions more trustworthy and verifiable in fields like medicine and law.
Researchers introduced DrawingVQA, the first benchmark to evaluate multimodal large language models (MLLMs) on real-world construction drawings — a uniquely complex domain fusing abstract geometry, symbols, tables, and technical text. The benchmark uses 33 professional 'Issued for Construction' drawings and 92 expert-crafted questions to test AI's visual-textual reasoning in architecture and civil engineering.
Researchers introduced GraphDx, a multi-agent AI framework that uses a Medical Diagnosis Knowledge Graph to balance diagnostic accuracy with resource costs, reducing unnecessary testing in sequential medical diagnosis.
Cognikernel is a new open-source tool that gives AI coding assistants like Claude Code and Codex persistent local memory. It automatically captures architectural decisions and project context during coding sessions, eliminating the need for manual markdown files.
Researchers from the University of Washington and NYU Abu Dhabi introduced CAMMAR, a representation learning framework that organizes Arabic meanings into nested lexical, cultural, and metaphorical embedding subspaces. This approach addresses 'semantic smearing' in current Arabic language models and could improve translation, chatbots, and other AI tools for Arabic speakers.
The Model Context Protocol (MCP), a key building block for AI interoperability, is getting a simpler version that reduces the complexity and cost of connecting AI models to external data sources and services. This could lead to more seamless integrations in everyday apps and tools.
A new study on 4,181 math problems found that adding AI reviewers to multi-agent systems doesn't always improve accuracy. While reviewers help with harder problems, they don't boost results for easier ones. For the most difficult problems, a simple peer discussion method outperformed a structured planner-executor-reviewer pipeline.
A leaked demo from Google DeepMind shows an AI-powered search engine that understands complex, multi-part questions and generates detailed, structured responses — moving beyond link lists to direct answers, summaries, and comparison charts.
A new arXiv study proposes converting all patient data—clinical narratives, vital signs, lab values, and comorbidities—into a single natural language sequence and fine-tuning a pretrained LLM, eliminating the need for task-specific fusion architectures in clinical prediction.
AI data centers face a critical power shortage in 2026 as the rapid growth of AI models outpaces grid infrastructure. The bottleneck threatens to slow AI advancements, increase costs, and raise environmental concerns, particularly in high-density regions like Northern Virginia and the Pacific Northwest.
Bloomberg reports that AI-driven automation may reduce income tax revenue as fewer people pay taxes, potentially destabilizing economies. Governments will need to adapt their tax policies to address this shift.