Translate
Live intelligence

Latest AI updates

1,733 updates
AllModelsProductsIndustryResearchTutorialOpinion
Research Editor selected

Smartphone-Aided VR Exergames for HIIT: Real-Time Pose Estimation and AI Feedback

This paper presents a VR exergame system for HIIT that uses a standard smartphone camera for full-body tracking, processed via real-time pose estimation and analyzed by an MCP-based backend with GPT-4o. A 20-participant study found higher enjoyment, immersion, and recommendation likelihood in VR versus non-VR, with AI feedback agreeing with expert evaluation at 87.5%.

Why it matters: Offers a cost-effective, scalable approach to VR fitness with validated AI feedback, relevant to both fitness tech and AI applications.

arxiv2d agoScore 82
Virtual reality fitnessAI-driven feedbackPose estimation
Research

TailSteer: Trajectory-Tail Steering for Drift-Resistant Multi-turn Image Editing

TailSteer is a final-stage training framework that improves multi-turn image editing robustness by concentrating preservation corrections near the trajectory tail of the denoising process. It learns from the editor's own on-policy rollouts using unchanged source content as reference, avoiding curated ground-truth pairs. Experiments on FLUX, SenseNova, and Qwen families show reduced cumulative drift over 15-20 turns while preserving base-model performance.

arxiv2d agoScore 78
image editingmulti-turn robustnessfoundation models
Research

SARBLOH: Spatial Perception Engine Boosts ARC-AGI-3 Scores 13.7x

A new open-weight agent, SARBLOH, demonstrates that spatial perception, not planning, is key to interactive reasoning on ARC-AGI-3. Its zero-parameter perception engine converts raw grids into object inventories and geometric briefings, raising rubric scores from 3/28 to 25/28 and cutting token use by 89.4%. On hidden Kaggle evaluation, it scored 16.53, 13.7x the first Milestone winner, within a nine-hour budget.

arxiv2d agoScore 74
ARC-AGI-3spatial reasoninginteractive agents
Research

Message-Passing Graph Persistence (MGP) enhances GNN expressivity with topology-aware descriptors

The paper introduces Message-Passing Graph Persistence (MGP), a family of topological descriptors that integrates message passing into graph filtrations, updating node embeddings at each step and linking them to persistence events. MGP is proven to be invariant under graph isomorphism and strictly more expressive than existing PH-based descriptors. Experiments on expressivity benchmarks and graph classification/regression datasets show MGP outperforms other PH-based descriptors on most datasets.

arxiv2d agoScore 74
Graph Neural NetworksTopological Data AnalysisPersistent Homology
Research

Task-Efficient Agent Evaluation Using Cached Traces

This paper introduces a method to estimate an agentic system's benchmark performance from a small number of task executions by leveraging rich execution traces. The approach extracts structured summaries with an LLM and embeds them to create representations that preserve task-related signal while suppressing identity-related noise. On SWE-bench Verified and Terminal-Bench 2.0, combining this geometry-based prediction with an item-response model reduces mean absolute error significantly compared to baselines.

arxiv2d agoScore 75
AI evaluationagentic systemsefficient benchmarking
Research

Training and Calibration Data for Open Models on ARC-AGI-2

This paper investigates which training and calibration data help three open models solve ARC-AGI-2 tasks offline. The authors built a corpus of executable rules and synthetic tasks, but found that fine-tuning on real task records outperformed synthetic data. For a reasoning model, an inference procedure contributed most of the score gain, while expert pruning and routing statistics improved a mixture-of-experts model's performance.

arxiv2d agoScore 79
ARC-AGI-2open modelstraining data
Research

Metamodels Forecast LLM Answer Activations Before First Token

Researchers propose context-answer metamodels that map context activations to forecasts of mean-pooled answer activations, capturing variability across stochastic rollouts. The metamodels achieve high fidelity, scale with training data up to 500,000 prompts, and generalize zero-shot to out-of-domain datasets. Forecasts retain behavior-predictive signal and transfer across models, suggesting scalable pre-generation behavior prediction.

arxiv2d agoScore 75
language model behavior predictionactivation forecastingefficient evaluation
Research

ServerBench: Benchmarking Coding Agents for Designing Specialized LLM Serving Systems

ServerBench is a new benchmark that evaluates coding agents' ability to design complete, specialized LLM serving systems. Agents are scored on a dollars-per-dollar metric, balancing latency-qualified token value against GPU cost. In tests across 11 coding-agent models, Claude Sonnet 5.5 achieved 2.97 dollars per dollar on Gemma E4B, outperforming the vLLM reference by 4.87 times. The benchmark includes stateful replay and independent verification to ensure correctness.

arxiv2d agoScore 77
LLM serving systemsagentic software designbenchmarking
Research

CompressionBench: Benchmarking Coding Agents for Autonomous LLM Compression

CompressionBench evaluates whether coding agents can autonomously compress pretrained LLMs within storage and time budgets. Across six agents and six model sizes, 29 of 36 submissions passed verification, achieving a 3.4% perplexity increase over BF16 and 77.33% downstream accuracy. The study highlights representation design, experimental judgment, and capability validation as key factors in agent-driven compression.

arxiv2d agoScore 77
LLM compressionautonomous agentsrecursive self-improvement
Research

Turn-Level Best-of-N Sampling Boosts LLM Agent Performance

This paper introduces turn-level best-of-N sampling, a simple inference-time scaling method for LLM agents on long-horizon tasks. Using a generator-selector framework with two off-the-shelf open-weight LLMs, the approach improves accuracy by 5-10% and consistency across repeated attempts. Analysis shows agents adopt desirable behaviors like fewer attempts, better instruction following, and improved debugging.

arxiv2d agoScore 72
LLM agentsinference-time scalingsampling methods
Research Editor selected

Occlusion for Free: Detecting Occlusion from the Disagreement Between Two Monocular Distance Cues

The paper presents a method to detect occlusion using the disagreement between two monocular distance cues—apparent height and metric depth—that a camera-only pipeline already computes. It requires no training or occlusion labels and achieves high AUC scores on KITTI (0.909) and CARLA (0.819), outperforming distance alone. The approach is robust to domain shifts and does not rely on focal length, stature, or depth scale.

Why it matters: Offers a practical, label-free occlusion detection method that works across datasets and deployment settings, addressing a key gap in real-world robotics and autonomous driving.

arxiv2d agoScore 87
occlusion detectionmonocular depth estimationautonomous driving
Research Editor selected

HALO: Perceptive Humanoid Motion from Simulation

HALO is a humanoid autonomy system that synthesizes vision-language-motion data across a thousand procedurally generated homes, scaling with compute rather than human labor. It acquires a repertoire of physical skills such as sitting, navigating, and crawling, and orchestrates them for long-horizon tasks. Without any real-world training data, HALO transfers zero-shot to physical environments, achieving a 91% success rate across 360 trials in six unmapped real-world settings.

Why it matters: Presents a scalable simulation-to-real pipeline for humanoid robots that achieves high zero-shot transfer success, addressing the data bottleneck in humanoid autonomy.

arxiv2d agoScore 90
humanoid autonomysimulation-to-reality transferrobot skill learning
Research Editor selected

FrontierPhysics: A New Benchmark for End-to-End AI Physics Research

FrontierPhysics is a benchmark created by 42 physicists with 56 tasks spanning theoretical, computational, and experimental physics. It evaluates AI agents on the full research pipeline—literature review, planning, experimentation, and manuscript writing—using both deterministic verifiers and expert rubrics. Across 22 models in 26 configurations, no agent passed more than 46 of 168 rollouts, with the main weaknesses being insufficiently rigorous reasoning and limited expert intuition rather than knowledge gaps.

Why it matters: Provides a rigorous, expert-curated evaluation of frontier AI agents' ability to conduct end-to-end physics research, highlighting critical reasoning deficiencies.

arxiv2d agoScore 89
AI agentsphysics researchbenchmark evaluation
Research Editor selected

BEAKER: An Expert-Curated Benchmark for Embodied Brains in Self-Driving Chemical Laboratories

BEAKER is a new benchmark with 1,000 expert-curated QA samples across 11 task types to evaluate multimodal LLMs as embodied brains in self-driving chemical labs. It tests affordance grounding, trajectory planning, equipment/safety recognition, and procedural reasoning. Evaluations of 28 MLLMs reveal that while models grasp coarse procedures, they struggle with fine-grained action grounding and precise perception. The benchmark highlights a critical gap for future embodied AI in laboratory automation.

Why it matters: Provides a rigorous, expert-curated diagnostic benchmark that exposes specific weaknesses in current MLLMs for embodied chemical lab tasks, guiding future research.

arxiv2d agoScore 89
embodied AImultimodal large language modelsself-driving chemical laboratories
Products Editor selected

GPT-Image-2.6全量上线：直出GIF和SVG，6.2 Astra同步现身

OpenAI发布GPT-Image-2.6，支持直接生成GIF和SVG格式，最快6秒出图，并同步推出6.2 Astra模型。该版本在图像生成速度和格式支持上实现突破，为创意工具和开发工作流带来新可能。

Why it matters: 该更新涉及具体模型版本和功能增强，对AI图像生成领域有直接实用价值。

qbitai2d agoScore 85
OpenAI图像生成模型更新
Industry Editor selected

OpenAI-funded audit finds 948 vulnerabilities in Linux kernel

An audit funded by OpenAI at a cost of $1.2 million uncovered 948 vulnerabilities in the Linux kernel. The findings highlight significant security gaps in the widely used open-source operating system. The audit's scale and funding underscore growing concerns about kernel-level security.

Why it matters: Provides concrete, recent data on a large-scale security audit with direct AI-industry involvement, valuable for readers tracking AI infrastructure security.

qbitai2d agoScore 85
Linux securityAI infrastructureOpenAI funding
Industry

Apple reportedly cutting budgets for internal AI usage

Apple is reportedly significantly reducing spending on AI for its engineering teams, according to TechCrunch. The move suggests a shift in internal AI investment priorities. Details on the exact budget cuts or affected projects are not provided.

TechCrunch AI2d agoScore 78
Apple AI strategycorporate AI spendingAI investment
Research

How scientists use Claude: Anthropic study on AI adoption in research

Anthropic published a report using privacy-preserving infrastructure to analyze how scientists use Claude, Claude Code, and Claude Science. The study documents usage variation across scientific fields and research tasks, identifies drivers of these patterns, and outlines scenarios for AI-accelerated research. It provides empirical insights into real-world AI adoption in academic and industrial research settings.

AIHUB2d agoScore 75
AnthropicClaudeAI in science