Translate
Live intelligence

Latest AI updates

1,729 updates
AllModelsProductsIndustryResearchTutorialOpinion
Research Editor selected

Dynamical Memory Transformer: Selective Recurrent Memories with Constant Cost Attention

Researchers present the Dynamical Memory Transformer (DMT), a hybrid architecture that separates sequence encoding from information reading. Several memory units, each a dynamical system with a complex diagonal recurrent matrix and an input-dependent leak rate, are read by attention without going through token history. The recurrence supports an associative scan with O(T) total work and O(log T) parallel depth, while recurrent states and the convolution cache remain fixed-size during autoregressive inference. Experiments show gains on several cognitive tasks and stronger length extrapolation on selective copy, with language-modeling performance close to Transformer but behind Mamba.

Why it matters: It offers a concrete architectural alternative for constant-cost long-context inference, with candid comparisons against Transformer and Mamba.

arxiv5d agoScore 82
Dynamical Memory TransformerEfficient attentionLong-context modeling
Research

SimXAI: A Lightweight LLM-Centric Layer Improves Conversational XAI

SimXAI replaces the rigid GUS-style modules of conversational explainable AI agents with a single tool-calling LLM that bridges to existing XAI resources. It achieves strong function-selection accuracy and high faithfulness while better handling context-dependent follow-up questions. In a user study replacing only the conversational layer of TalkToModel, users completed tasks more accurately and in fewer turns.

arxiv5d agoScore 78
explainable AIconversational agentsLLM integration
Research

LIT: Co-Evolving Video Director Agents Improve Production and Interaction

The paper introduces LIT, a framework for video director agents that co-evolves directing and user interaction capabilities. It uses a simulated user to probe weaknesses and a meta engineer to revise strategies, forming an evolutionary loop. On Test and OOD Test splits, LIT improves production and interaction scores by 19.1% and 32.0% relative to the initial pipeline.

arxiv5d agoScore 72
AI agent evolutionvideo generationhuman-AI interaction
Research Editor selected

AI evaluators favor job applications written by certain AI models, independent of candidate quality

A new study introduces a multi-model auditing framework that pairs eleven LLM writer agents with nine evaluator agents across ten job descriptions and fifty candidates per job. It finds a consistent generative preference hierarchy where locally hosted models are systematically disadvantaged, compounded by self-preference bias. In the most adverse pairing, about 1.8 out of 25 qualified candidates are filtered out purely due to model mismatch. The study offers actionable guidance for fairer screening design and discusses implications for algorithmic fairness and regulation.

Why it matters: This research provides concrete evidence of AI model bias in hiring, a high-impact area with practical implications for fairness and regulation.

arxiv5d agoScore 85
algorithmic fairnessAI-assisted screeningLLM evaluation
Research

Automated Keyword Generation Using LLMs in a High-Volume Data Virtualization Service

This paper presents a solution for automatically generating keywords for virtual datasets in a data virtualization service (DVS) using large language models (LLMs). The authors propose five strategies for selecting text data for LLM processing and evaluate them across three LLM generators and two embedding techniques in five application case studies. Results show that strategy performance depends on text description quality, leading to a multi-strategy implementation in a new keyword generation service. The work highlights the importance of automated keyword generation for dataset-rich, frequently accessed DVS environments.

arxiv5d agoScore 72
large language modelsdata managementautomated metadata generation
Research Editor selected

General-purpose foundation models outperform specialized chemical FMs on property prediction with scarce data

A new arXiv study compares specialized chemical foundation models (ChemBERTa, MoLFormer) with general-purpose models (GPT-6 Astra, Claude Opus 5.5) on chemical property prediction tasks. With only 60 experimental datapoints, GPT-6 Astra outperforms the chemical FMs on five regression tasks. Chemical FMs only match Astra's performance when given 10x more data, suggesting general-purpose models are better for data-scarce chemistry research.

Why it matters: This finding challenges the common assumption that domain-specific models are superior in specialized fields, offering practical guidance for researchers in chemistry and AI.

arxiv5d agoScore 82
foundation modelschemistryproperty prediction
Research

ArchOR: Structure-Aware Skill Retrieval Improves LLM Optimization Modeling

ArchOR is a new archetype-centric skill agent for LLM-based optimization modeling, building on OptSkills with two retrieval modules: Centroid-Aware Representation Learning (CARL) and Signature-Guided Retrieval (S3R). It achieves 65.00% Micro-Avg. Pass@1 across five benchmarks, improving over OptSkills by 1.64 pp Micro-Avg. and 2.68 pp Macro-Avg. The method uses structural signatures and a baseline-preserving arbitration rule to enhance robustness under narrative rewrites.

arxiv5d agoScore 72
large language modelsoptimizationretrieval
Research Editor selected

Observable-Specific Validity Tests for Scientific AI

This paper introduces observable-specific validity records for scientific AI, linking reductions to reference models, error measures, tolerances, parameter domains, and interventions. It demonstrates explicit certificates for a singularly perturbed oscillator and Hubbard dimer relations, and shows that imposing dimensional homogeneity reduces energy scale transfer error from 52.3% to 0.0768% across 16 neural fits. The work also derives an exact fixed control query requirement for mechanism discrimination and provides a 256-task evaluation interface.

Why it matters: Provides a concrete, testable framework for validating scientific AI models, with quantitative results and practical implications for physical transfer and mechanism discrimination.

arxiv5d agoScore 82
scientific AI validationcontrolled approximationphysical transfer
Products

Sensorium: Open-Source Library for Multimodal Surgical Data Collection

Sensorium is an open-source Python library for recording, processing, annotating, and curating multimodal surgical data from robot and sensor streams using ROS/ROS2. It was evaluated across 1516 recordings from phantom, ex vivo, and in vivo porcine experiments. The library aims to reduce cost and improve repeatability in embodied surgical AI data collection.

arxiv5d agoScore 72
surgical roboticsdata curationmultimodal learning
Research

Geo-FSRL: Geometry-Optimized Few-Step Flow Reinforcement Learning

Geo-FSRL introduces a selection principle for stochastic policies in few-step flow models, minimizing conditional KL divergence for velocity-output updates. It derives a closed-form optimizer within a coefficient-preserving family of isotropic Gaussian transitions, improving training stability. The method achieves higher final reward scores than Flow-GRPO and Flow-CPS on image and video generation under 4-step settings.

arxiv5d agoScore 78
few-step flow modelsreinforcement learninggenerative AI
Research

Aligning MLLMs with Human Cognition via Behavior-Derived Semantic Dimensions

Researchers propose a framework to align multimodal large language models with human cognitive structure by using a behavior-derived semantic embedding space of 66 dimensions from 4.7 million human Odd-One-Out judgments. The method infers salient decision dimensions for each triplet and generates dimension-grounded rationales, improving consistency with human similarity judgments and increasing representational alignment in fMRI data. The approach retains competitive performance on standard multimodal benchmarks.

arxiv5d agoScore 78
multimodal large language modelshuman-AI alignmentcognitive science
Research Editor selected

Chronos-2: A Foundation Model for Multivariate Time Series Forecasting

Chronos-2 is a new foundation model for zero-shot forecasting that models interactions among multivariate time series and heterogeneous covariates, unlike prior models that treat series independently. It achieves state-of-the-art accuracy on average, improving skill scores by 5.2 percentage points and win rates by 10 percentage points over strong baselines. The model is computationally efficient, generating forecasts for 100 series in under a second on a mid-range GPU, and matches or surpasses specialized methods in domains like infectious disease, macroeconomics, carbon flux, and runoff.

Why it matters: This introduces a significant advance in time series foundation models, enabling multivariate interaction modeling and broad cross-domain applicability with strong performance and efficiency.

arxiv5d agoScore 85
Chronos-2time series forecastingfoundation models
Research Editor selected

ETAPS: Gradual Pruning with Sparse Adaptation Achieves High Quality at Extreme Sparsity

Researchers propose ETAPS, a pruning method that gradually tightens a global sparsity constraint while training a sparse delta with a small fixed budget. This approach avoids model-sized optimizer states, enabling efficient pruning of LLMs at extreme sparsity levels. Across models from 125M to 13B parameters, ETAPS achieves lower perplexity than the prior state-of-the-art ELSA at sparsities from 50% to 90%, with up to 42% perplexity reduction. It also reduces cost significantly, pruning LLaMA2-7B on a single H200 with 5x smaller persistent state and 49% fewer GPU-hours compared to ELSA.

Why it matters: This paper presents a concrete, recent advancement in LLM pruning that improves both quality and efficiency, offering practical value for deploying large models.

arxiv5d agoScore 85
model compressionlarge language modelsefficient inference
Research

Audit Finds Test-Time Search Gains Limited for World Action Models

A pre-registered audit of test-time scaling in world action models on Cosmos Policy and RoboCasa reveals that learned value selectors underperform random selection, and oracle gains are minimal. Token pruning reduces FLOPs but fails to achieve target speedups, and a silent rendering corruption was identified and fixed. The study emphasizes auditing test-time scaling as a chain of premises rather than relying on success rates alone.

arxiv5d agoScore 78
test-time scalingworld action modelsAI evaluation
Research

ID2Pattern: Pretraining Generalizable Identity Behavior Patterns for Transaction Risk Foundation Models

ID2Pattern is a pretraining framework that captures identity behavior patterns in transaction sequences using a Local Anonymous ID tokenizer and a multi-scale pattern codebook. It outperforms strong baselines on an industrial dataset with over 100M sequences and six public datasets, improving performance by 4.13% and 8.83% on average respectively. The method remains effective even when identity values are unseen during training, showing potential for generalization to new users and markets.

arxiv5d agoScore 78
transaction risk modelingbehavioral pretrainingidentity encoding
Research Editor selected

DAMI: Dual-Axis Memory Initialization Boosts Long-Context Retrieval in Gated Linear Attention

DAMI is a new initialization method for gated linear attention models that configures gate parameters and input activations before training to improve long-context retrieval. It assigns geometric timescales across heads and a low-high-low input-sensitivity profile across layers, calibrated with a small activation sample. Across five model families, DAMI improves RULER-style retrieval accuracy by 5.29 to 16.71 percentage points over random initialization baselines, while keeping language modeling and reasoning performance comparable.

Why it matters: This paper presents a practical, training-free initialization technique that yields significant retrieval gains across multiple state-of-the-art linear attention models, offering immediate value for researchers and practitioners working on long-context AI.

arxiv5d agoScore 82
Gated Linear AttentionLong-Context RetrievalModel Initialization
Research Editor selected

SecPO: Adversarial Training and New Loss Cut Prompt Injection Attack Success Rates Dramatically

Researchers propose SecPO, a principled adversarial training pipeline for defending LLM agents against prompt injection attacks. They introduce a SecPO loss that approximates an idealized injection-invariant reference model, and combine it with efficient synthesis of strong injected inputs. Their SecPO + AT model reduces attack success rates by up to 14.5x compared to Meta-SecAlign, and further 2.4-4.6x over DPO loss, while maintaining utility across nine benchmarks.

Why it matters: This paper presents a concrete, state-of-the-art defense against a critical AI security threat, with significant measured improvements and released resources.

arxiv5d agoScore 85
AI securityprompt injectionadversarial training
Research

AutoAnalysis: Automated Self-Updating Meta-Analysis of Human-AI Collaboration

AutoAnalysis is an automated, self-updating meta-analysis that synthesizes evidence on when humans and AI work better together or alone. Its latest run includes 1,955 effect sizes from 420 papers, finding that human-AI combinations outperform humans alone but underperform the better of either alone. The system integrates new studies as they emerge, providing cumulative evidence across tasks and AI agents.

arxiv5d agoScore 78
human-AI collaborationmeta-analysisAI performance evaluation