New guidance highlights kidney risks in adult IgA vasculitis
The peer-reviewed framework offers 14 recommendations, while acknowledging that treatment evidence remains limited.
Search
2,111 results
201–220
Weight-bearing knee test detects proprioception gaps after ACL surgery
A peer-reviewed study found the Knee-VEDA test was reliable overall and distinguished operated limbs from healthy controls, but it needs more validation.
Sulfuric-acid biochar linked to stronger sugar beet growth in saline soil
A seedling pot experiment found the sulfuric-acid treatment had the strongest reported gains, but its short-term design does not establish field performance.
Fatal child brain injuries show sharp age and mechanism patterns
A retrospective study found different injury profiles in traffic accidents, falls and abusive head trauma, but the results do not establish cause.
Adults with type 1 diabetes wore insulin sets longer in six-month study
A prospective, non-randomized study reported longer wear and higher satisfaction, but only 48.3% maintained the full seven-day duration.
Gene study finds shared aging signals in periodontitis and rheumatoid arthritis
A computational analysis identified five shared hub genes and strong model performance, but the findings still need clinical and laboratory validation.
Survey finds hurdles to pancreatic cancer prehabilitation
Healthcare professionals backed shared responsibility and tailored, earlier support, but flagged costs, symptoms and information overload as obstacles.
Tensor constructions lower theoretical matrix-multiplication bounds
An arXiv preprint dated 27 August 2026 reports exact centroid formulas, conditional rank theorems and lower laser-method upper bounds, without measured algorithmic performance.
Proposed AI framework keeps clinicians in charge of medication decisions
Peer-reviewed conceptual work outlines seven design principles for a hybrid system that combines evidence retrieval, safety rules and language-model reasoning.
Sacramento study will test whether cool water becomes routine burn first aid
A peer-reviewed protocol will compare 774 patient-record reviews before and after implementation while asking clinicians, burn survivors and caregivers whether the approach fits care.
AI agents look beyond top hotel listings, but rank still matters
Preprint: In a simulated Manhattan hotel search, AI agents inspected more options than a human benchmark, while rank effects on choices varied by model and reasoning effort.
AI Recovers Muscle and Motion Clues From 3D Tongue Meshes
Preprint: A simulator-grounded dataset tested models on synthetic tongue geometry, with steep score drops when meshes were mismatched.
AI system reports more targeted, less toxic counterspeech
Preprint: FIRE reported higher factuality and category scores, higher human win rates across four criteria, and a 75% lower peak memory requirement in benchmark tests.
New AI prompting method reports gains in multimodal humor tests
Preprint: A code-based reasoning graph was associated with higher benchmark scores, but the study did not establish a causal advantage or human-level humor understanding.
Cybersecurity AI Leads Benchmark With 58.1% CyberGym Score
Preprint: OpenAegis reported the highest score among four models in a one-hour-per-task test, while the authors say broader evaluation is still needed.
Quantum oscillator keeps one central peak in a double well
Preprint: An exact model finds a single-peaked ground state above the internal barrier, while changing representation alters its visible quantum structure.
Graph-Based Optimiser Reports Zero Final Error on Selected Problems
Preprint: Mycelial Search matched other optimisers on selected CEC 2022 functions, but its results varied with landscape structure and information flow.
Fisher-guided quantization reports less quality loss at low bit rates
Preprint: The system reported lower perplexity and higher benchmark accuracy than several baselines near three-bit compression, but AWQ was faster on the tested GPU.
A trading framework says average backtests are not enough
Preprint: A conditional model ties tail risk, search costs and position sizing to its assumptions, but its one-index test found no robust position.
Calcium Isotopes Could Reveal Nuclear Shells in Collisions
Preprint calculations suggest that coherent J/ψ production could make shell filling visible in diffractive patterns, while ϕ production is more sensitive to saturation effects.