Search the full wire by company, model, lab, or keyword. Every story we have ever aggregated.
Mental World Model framework for simulating subjective cognition and Theory of Mind extends LLMs beyond linguistic inference with dynamic evolution.
LLM medical verification fails to distinguish independent trials from same-trial republication, revealing aggregation brittleness across five model conditions.
Neural Expansion framework explains deep network generalization via ray coverage geometry, unifying margin bounds and neural tangent kernels mechanistically.
RoBERTa-RT model measures cognitive load of LLM-generated vs. human text via reading time but findings are psycholinguistic not AI-architectural.
GLOBE agent framework enables self-improving e-commerce product optimization via structured agentic memory across categories and search trajectories.
Tool Recommendation system deployed for multi-step agents surfaces 8,200 tools through fixed interface, solving context-window bottleneck at scale.
Amulya Vishwanath, who spent seven years leading product marketing for NVIDIA’s robotics platforms, says robotics is currently in its “Lego era,” with the basic building blocks like actuators, robotic hands, software, and safety standards still being rebuilt. This view runs counter to those who believe that robotics, powered by a general foundation model, is about to have […]
Blitz distillation reduces AlphaFold3 and Boltz-2 denoising from 200 to 8 steps via geometric noise schedule and multistep consistency training.
VISTA infers latent user values from survey responses and uses hypernetworks to steer frozen LLMs toward personalized outputs without prompt engineering.
Statistical protocol using two-level bootstrap and TOST for certifying practical equivalence in ML experiments with rigorous power theory.
Theoretical analysis of IQP quantum kernel expressivity via RKHS characterization for quantum machine learning feature maps.
HardEval benchmark measures LLM/agent capability to implement STIG security controls on live systems with persistent state changes.
Comprehensive evaluation of Qwen3-TTS zero-shot voice cloning across interfaces, domains (read/meeting/emotional/ATC speech), and failure modes vs. XTTS-v2 and F5-TTS.
Framework assessing reliability of LLM agents simulating stakeholders for multi-criteria healthcare operations decision-making under preference uncertainty.
OwlSight introduces Factor-based CoT monitoring to detect and oversee reasoning models across diverse tasks with customizable oversight requirements.
according to Elorian's evaluations, Glimpse outperforms even GPT-6 Astra on scientific chart reasoning and short-form video captioning, with competitive results on other visual reasoning tasks.
TUPLE-Bench evaluates privacy leakage from confidence scores and probability distributions in typed classifiers; shows 91-99% exposure despite instructions to withhold.
TriRouter: electrode-space routing framework for robust EEG motor imagery decoding under sparse or corrupted multichannel sensor configurations.
YOLOv8n model for automated banana surface damage classification to reduce manual inspection and spoilage detection in agribusiness.
LILIE uses vision-language models with normalizing flows to supervise visual generation via caption recovery, replacing pixel-space loss.
Case study examining AI's role in contemporary musical composition through two electroacoustic works and ethical implications.
JAXtari accelerates Arcade Learning Environment via JAX with GPU support and object-extraction to test agent robustness vs. visual heuristics.
ML interatomic potentials for materials under electric fields via analytic long-range electrostatics separation enabling distributed computation.
No-regret generation framework with tree-based mistake-forcing dimensions characterizing minimax complexity for sequence-level language generation.
EEG decoder analysis across Llama-3.1-8B layers reveals passage-level context structure in silent reading aligns with deep layer embeddings.
Meta “streamlined” its commercial TOS. Legal experts warn it’s an AI power grab.
The data comes from Vercel's AI gateway, a unified API service similar to OpenRouter.
PresenceBench benchmark measures LLM adaptation to user functional state (readiness, flow, stability) via receptivity and specificity metrics.
Multi-task active learning framework using orthogonal decomposition instead of linear scalarization to resolve gradient conflicts across objectives.
GRAFT-EEG enables cross-dataset motor-imagery EEG decoding via gauge-aware physical field transport without target statistics or labels.