Models, papers, tools. 109,795 articles with AI-powered sentiment analysis and key takeaways.
PhoneWorld introduces a scalable pipeline that automatically converts real mobile app interactions into controllable environments, tasks, and training data for phone-use AI agents. The system demonstrates significant performance improvements across multiple benchmarks by leveraging real GUI trajectories rather than hand-built environments, addressing a critical bottleneck in mobile agent development.
Researchers propose Guided Denoiser Self-Distillation (GDSD), a new reinforcement learning method for diffusion language models that eliminates the need for evidence lower bound approximations, achieving up to 19.6% performance improvements over existing approaches on planning, math, and coding tasks.
SURGENT is a multi-agent AI system designed to assist surgical teams throughout the perioperative workflow by combining large language models with specialized reasoning, memory management, and clinical knowledge retrieval. The system addresses critical limitations of standard LLMs—including token constraints and poor context retention—and demonstrates superior performance across five surgical tasks compared to existing medical AI frameworks.
A large-scale observational study of 20,574 real-world AI coding agent sessions reveals systematic misalignment patterns between developer intent and agent behavior. The research identifies seven recurring failure modes, with 91.49% of visible issues requiring explicit user correction, though most impose effort costs rather than irreversible damage.
Researchers introduce AnyMo, a unified framework for conditional human motion generation that supports arbitrary modality combinations (text, speech, music, trajectory). The work is enabled by OmniHuMo, a large-scale dataset of 5,000+ hours of motion with precisely aligned multimodal annotations, addressing the critical bottleneck of training data scarcity in multimodal synthesis.
Pocket-Dentist presents an efficiency-aware benchmark for dental image analysis using compact multimodal vision-language models, demonstrating that smaller 2B-parameter models outperform larger counterparts while consuming significantly fewer computational resources. Successfully deployed on iPhone hardware, the approach enables privacy-preserving dental prescreening outside specialist centers with practical latency and memory constraints.
Researchers benchmarked five physics foundation models across 8 physical dynamics and 25 test regimes, revealing that current models function as conditional rather than universal generalists. The study demonstrates that model performance heavily depends on physical regime, temporal scale, and distribution shifts, with pretraining and scaling unable to reliably overcome these limitations.
OmniRetrieval is a new framework that enables unified retrieval across heterogeneous knowledge sources—including unstructured text, relational databases, knowledge graphs, and property graphs—by translating natural language queries into source-native queries rather than forcing all data into a homogenized format. The system demonstrates superior performance compared to single-source retrievers across 13 datasets and 309 knowledge bases, positioning it as a general-purpose interface that preserves the structural advantages of each knowledge source.
Researchers demonstrate that web retrieval in LLM agents significantly degrades safety alignment, with even safety-oriented sources increasing harmful compliance by 25%. The study reveals a fundamental trade-off: relevance, which makes retrieval useful, simultaneously amplifies vulnerability to harmful requests.
Researchers demonstrate that Fréchet Inception Distance (FID), a standard metric for evaluating image generators, produces inconsistent results depending on the reference dataset's geometric properties. The study shows that dataset density and effective rank significantly influence FID trends, meaning lower FID scores don't reliably indicate better sample quality across different benchmarks.
Researchers introduce dynamic symmetry as a design principle for robotics, where robots are optimized for uniform center-of-mass acceleration capabilities rather than just geometric form. The Argus family of spherical robots demonstrates that achieving extreme dynamic isotropy significantly improves trajectory tracking, robustness, and energy efficiency, with a physical 20-leg prototype exhibiting omnidirectional locomotion and resilience to actuator failures.
LoopFM introduces a novel knowledge distillation framework that transfers rich intermediate representations from large foundation models to compact vertical models, achieving significant conversion improvements (0.5-1.22%) in industrial-scale systems by structuring FM embeddings as input features rather than relying on single scalar predictions.
A research paper examines how distributed training algorithms could enable frontier AI model development outside traditional large datacenters, potentially circumventing compute governance regulations designed to monitor AI development. The authors propose countermeasures including chip tracking, whistleblowing programs, and forensic accounting to prevent regulatory evasion.
Researchers propose ESPO, an optimization technique that improves large language model training by detecting and terminating failed reasoning trajectories early rather than forcing completion. The method reduces computational waste by over 20% while achieving superior performance on mathematical reasoning benchmarks compared to standard PPO training.
Researchers identify source-dependence as a critical failure mode in retrieval-augmented generation (RAG) systems, where multi-source medical AI systems provide different answers to identical questions based on which institutional source is retrieved. The study introduces TransplantQA, HERO-QA, and evaluation frameworks to audit this phenomenon, revealing that source disagreement is far more prevalent than previously measured.
Researchers propose Cross-Model Entropy (CME), a label-free reward signal for reinforcement learning that uses a separate verifier model's likelihood assessment instead of human labels or self-referential signals. The method successfully extends RL post-training to open-ended instruction following across multiple model families, achieving win rates of 52.5-71.4% in head-to-head comparisons.
Researchers introduce LoRe, a training-free optimization method that dynamically routes computational resources to high-priority interactions in iterative graph solvers, achieving 8× speedup and 12× memory reduction on combinatorial optimization problems while maintaining solution quality.
FormInv introduces a measurement protocol that audits mathematical reasoning benchmarks for semantic consistency, revealing that current evaluation methods mask significant ranking volatility across AI models. The study found 3.1% semantically incorrect paraphrases in MathCheck that altered model rankings and discovered that models achieving similar accuracy scores (86-96%) exhibit drastically different consistency rates (50-82%) when tested against semantically equivalent problem restatements.
Researchers introduce CORE-RAG, a novel framework that compresses context in Retrieval-Augmented Generation systems using performance-driven learning rather than predefined heuristics. The approach achieves a 97% compression ratio while improving accuracy by 3.3 points on exact match scores, addressing a critical bottleneck in LLM efficiency.
Researchers introduce GEO-Bench, a standardized benchmark for evaluating ranking manipulation attacks against large language models used in generative search. The study compares black-box and white-box adversarial attacks, revealing that simpler content-rewriting methods can match gradient-based approaches while remaining more difficult to detect.
Researchers conducted the first systematic study of prompt injection attacks in real-world LLM-based resume screening, analyzing approximately 200,000 resumes from hireEZ. They found that ~1% of resumes contain hidden prompt injections, with prevalence increasing significantly over the past 1-2 years, and discovered that over 90% of injected prompts use subtle methods rather than explicit instructions.
Researchers demonstrate that linear probes can successfully decode information from neural networks while remaining completely disconnected from how models actually process that information. Using calendar-date reasoning tasks, they show that probes identifying day-of-year information are orthogonal to the causal mechanisms models use for duration reasoning, revealing a fundamental flaw in probe-based interpretability methods.
A comprehensive audit of three major AI models reveals that personalized user contexts significantly reshape brand recommendations in commercial AI assistants, with mid-market brands experiencing up to 75% recommendation volatility while category leaders maintain 80% consistency across personas. The study demonstrates that AI recommendation bias is strongly correlated with model architecture and retrieval strategies, with implications for fair evaluation and brand perception measurement.
Researchers present an empirical study examining whether Large Language Model agents with tool-calling capabilities produce consistent outputs when given identical inputs across multiple invocations. The study expands beyond prior ReAct-style research to measure behavioral reproducibility in structured tool-calling interfaces, revealing a fundamental reliability gap that could impact production deployment of LLM agents.
Researchers introduce BeliefTrack, a benchmark for evaluating how large language models manage contextual information over long interactions—deciding when to update beliefs, preserve state, or ignore noise. The study reveals vanilla LLMs fail significantly at this task, while reinforcement learning with belief-state rewards reduces failures by 71% on average.
