Real-time AI-curated news from 89,338+ articles across 50+ sources. Sentiment analysis, importance scoring, and key takeaways — updated every 15 minutes.
RunAgent has developed SuperBrowser, an autonomous web navigation agent that mimics human browsing behavior through selective perception and structured memory management. The system achieves 89.47% success on the Mind2Web Hard benchmark, outperforming all published open-source baselines by applying consistent cognitive principles throughout its architecture.
A new study demonstrates that pairwise comparison methods like Elo, commonly used to evaluate generative AI models, produce rankings that correlate strongly (>0.9 Spearman correlation) with ground-truth accuracy benchmarks. The research shows these comparative evaluations substantially outperform direct judging when evaluators are weak and are largely resistant to stylistic bias and judge preference, though minor effects like answer repetition can influence outcomes.
Researchers found that structured output formats like JSON degrade AI model performance not because of formatting itself, but because of insufficient model capacity. Models with adequate computational headroom handle JSON constraints without accuracy loss, while smaller models operating near their limits suffer 28-36 percentage point drops, a penalty that can be partially recovered by reasoning first and formatting afterward.
Researchers propose Bayesian Selective Latent Inference (BSLI), a machine learning method that uses wastewater surveillance data to monitor influenza spread in communities before clinical cases are reported. The system intelligently decides whether additional data sources are needed or if abstention is appropriate, improving disease monitoring accuracy while managing data acquisition costs.
Researchers introduce TheoremBench, a comprehensive Lean4 benchmark for evaluating large language models on formal mathematics theorem proving. Unlike existing competition-focused benchmarks, TheoremBench assesses how LLMs handle longer, dependency-rich mathematical proofs through both standalone theorems and structured families of related subtasks, revealing that current models remain inefficient and biased toward simpler problems.
Researchers demonstrate that finetuning large language models on narrow safety tasks can induce broad alignment improvements—the opposite of previously documented emergent misalignment. Using Constitutional AI with four ethical frameworks (deontology, consequentialism, virtue ethics, and human authority), they show models develop consistent 'ethical personas' that generalize beyond their training data, though projectability varies significantly across approaches.
Researchers developed an LLM-orchestrated framework that automates conformance checking in healthcare by extracting patient care pathways and clinical guidelines from unstructured text, eliminating the need for formal Computer-Interpretable Guidelines. Testing at Alessandria Hospital's neurological ward showed 86% of stroke care traces adhered to clinical guidelines, demonstrating practical feasibility of AI-driven healthcare compliance assessment.
Researchers present MedSci Skills, an open-source toolkit that pairs LLM-assisted clinical manuscript generation with deterministic verification gates to detect fabricated citations, numerical errors, and missing reporting guidelines. The system demonstrates 100% detection of seeded defects versus 41% for generic LLM reviewers, providing an auditable trail for biomedical publishing.
A systematic literature review examines Self-Explainability (SX) in self-adaptive and self-organizing systems, finding that most approaches remain theoretical with no standardized evaluation methods. The research establishes a taxonomy and framework for advancing SX, identifying a significant gap between conceptual work and practical implementation in increasingly complex AI-driven systems.
Researchers introduced TABVERSE, a new benchmark for evaluating how Large Language Models and Vision-Language Models understand tables across different formats (HTML, Markdown, LaTeX, and images). The study reveals that table representation significantly impacts model performance, with structured text formats generally outperforming rendered images, though performance varies by task and model type.
Researchers propose an operational framework for evaluating recursive self-design in AI systems, where AI assists in modifying its own development mechanisms. The paper maps existing systems against four criteria and reports that Darwin Goedel Machine achieved significant performance improvements (20% to 50% on SWE-bench, 14.2% to 30.7% on Polyglot benchmarks) through iterative self-improvement over 80 cycles.
Researchers introduce CFips, a sampling algorithm for efficiently exploring interval patterns under user-defined constraints. The approach preserves exact sampling guarantees while decomposing syntactic constraints into elementary predicates, enabling pattern mining tasks that previously exceeded computational time limits.
Researchers demonstrate that pretrained biomedical language models fail catastrophically at cross-domain discrimination, assigning high similarity scores (0.76-0.92) to unrelated concepts. They propose BODHI, a contrastive learning approach that improves domain separation 2.3x while maintaining correlation accuracy, and show that optimized inference achieves 133x latency reduction on specialized hardware.
Researchers present Trellis, an autoformalization system that uses LLM agents within constrained workflows to convert natural language mathematical proofs into Lean formal code. The system achieves reliable formalization on modest computational budgets by enforcing incremental progress through iterative refinement, demonstrated by formalizing a recent Ramsey theory breakthrough.
Researchers present SearchSwarm, a framework that trains large language models to intelligently delegate complex tasks to subagents while managing finite context windows. The resulting 30B-parameter model achieves state-of-the-art performance on research benchmarks by learning when and what to delegate, addressing a critical bottleneck in agentic AI systems.
Researchers evaluate whether deep research agents (DRAs) can improve iteratively through feedback, finding that self-reflection yields negligible gains while single rounds of process-level feedback produce substantial improvements—but these gains don't compound over multiple turns due to regression on previously satisfied criteria.
Researchers introduce CHAP (Collaborative Human-Agent Protocol), a standardized framework for managing interactions between humans and AI agents in production systems. The protocol structures oversight moments, handoffs, and approvals as auditable events with cryptographic signatures, addressing a gap between existing tool-access standards (MCP) and agent-to-agent protocols (A2A).
Researchers introduce Evaluation Cards, a standardized reporting framework that addresses fragmented AI evaluation practices across leaderboards and model cards. The system consolidates benchmark metadata, evaluation data, and model information into unified records with interpretive signals for reproducibility and comparability, deployed across 5,816 models and 635 benchmarks.
Researchers introduce XAInomaly, an explainable AI framework using a Semi-supervised Deep Contractive Autoencoder for detecting anomalies in Open RAN (O-RAN) networks. The system addresses the critical need for interpretable machine learning in complex wireless infrastructure by combining generative modeling with explainability techniques to identify network traffic deviations while maintaining transparency in decision-making.
A research paper proposes blockchain as foundational infrastructure for embodied AI systems, addressing the dual challenge of securing data economies while defending against quantum computing threats. The work integrates post-quantum cryptography, cross-organizational governance, and scalable architectures to create trustworthy decentralized environments for AI-driven cyber-physical systems.
Researchers introduce a bidirectional search task linking code snippets with text descriptions and vice versa, addressing the gap between scientific publications and their implementations. They present a large dataset with automatically-generated training data and manually-annotated test sets, along with a modular encoder-based approach that achieves strong in-domain results with promising out-of-domain generalization.
Researchers investigating hallucinations in fine-tuned Large Language Models found that domain adaptation via fine-tuning alone is insufficient to prevent inaccurate outputs. Testing Llama-2 with domain-specific data revealed the model struggles with novel reasoning tasks and tends to over-generate information, highlighting fundamental limitations in current LLM adaptation techniques.
Researchers have successfully developed the first Retrieval Augmented Generation (RAG) system for legal question answering in Nepali, addressing a critical gap in AI applications for low-resource languages. The system achieved 91% precision using BM25 retrieval and demonstrated 84% human-evaluated truthfulness, establishing a viable foundation for AI-assisted legal services in non-English speaking jurisdictions.
Researchers introduce ABLE, a framework that represents and compares large language models through gradient-based feature attributions rather than parameter analysis or output comparison. The training-free method achieves competitive performance on model comparison tasks across 239 open-source LLMs while providing theoretical stability guarantees.
Researchers develop a novel method for constructing implicit causal graphs from text by using large language models to infer intermediate causal events between observed cause-effect pairs. The study compares multiple approaches including chain discovery and iterative search processes, validated against a curated database of 1,560 scientifically verified causal relationships.
