AI News

Encoding Invisible Causation for Bridge Diagnostic Agents: Triple-Guided Retrieval-Augmented Fine-Tuning with QLoRA (arxiv.org)

Neural Feature Governance: Extending Atom Prevalence (arxiv.org)

Physically Constrained Federated Additive Models for O-RAN SLA-Risk Prediction (arxiv.org)

Molt: A Scalable PyTorch-Native Training Framework for Agentic Reinforcement Learning (arxiv.org)

Shallower ReLU Network Representations via Exact Linear Algebra (arxiv.org)

A Drift Stable Quantum Federated Learning for Intelligent Services (arxiv.org)

Adjustment Speed as a Safety Constraint for Nonstationary Reinforcement Learning (arxiv.org)

Multi-Horizon Consistency as Geometry: When Latent Dynamics Contract, and When They Do Not (arxiv.org)

Toward Goal-Agnostic Joint-Embedding Predictive Control of Partial Differential Equations (arxiv.org)

Quasi-Monte Carlo Initialization for Meta-Reinforcement Learning (arxiv.org)

Prompt as a Data Type: In-Database LLM Prompt Management and Rewriting (arxiv.org)

Deep Sigma Point Processes for RCS Modeling in Spaceborne SAR Imagery (arxiv.org)

Explainable quantum-compressed machine learning for complex fluid flows (arxiv.org)

Ordered Action Tokens for Visuomotor Policy Learning (arxiv.org)

Generative and multimodal AI for materials prediction and design: Progress, challenges, and perspectives (arxiv.org)

Computer Vision Based Neurology Brain Activity Rejection Architecture and Implementation (arxiv.org)

Do Modules Stay in Their Lane? Role Drift in Compound LLM Systems (arxiv.org)

Adversarial Prompts for Acceptance Collapse in Speculative Decoding (arxiv.org)

FBLayout: Optimizing Memory Layout for Efficient LLM Finetuning on Mobile GPUs (arxiv.org)

Local Synaptic Rules Can Implement a SIGReg Gradient Without Backpropagation (arxiv.org)

FrED: External Data Influence Estimation via Domain Knowledge Graph Grounding (arxiv.org)

Procedural Knowledge Is Not Low-Rank: Why LoRA Fails to Internalize Multi-Step Procedures (arxiv.org)

SCOPE and SCION: A Benchmark and an Auditable Reference Pipeline for Schema Induction and Fusion from Text (arxiv.org)

Spectral Flow Certificates for Depth-Aware Long-Range Propagation in Graph Neural Networks (arxiv.org)

Do emulated quantum circuits change what CNNs look at? Performance and explainability comparison in medical image classification (arxiv.org)

An Explainable FFT-Based Spatial-Frequency Fusion Framework for Deepfake Detection (arxiv.org)

Susceptible Reservoir Architectures for Regime-Conditional Volatility Forecasting (arxiv.org)

\k{appa}-LoRA: Condition Numbers Reveal Which LoRA Matrices Worth Updating (arxiv.org)

Complexity Bounds and Approaches to Learning Projected Gradient Descent Solver Iterates (arxiv.org)

Phylogenetic signal in marine mammal and bird vocalizations captured by audio foundation models: the limited benefit of domain-specific pretraining (arxiv.org)

Hyperball May Not Be a Free Lunch (arxiv.org)

On the Identifiability of Controlled World Models (arxiv.org)

Relaxed activation analysis of dataflow networks - A clock calculus for machine learning and real-time scheduling (arxiv.org)

LunarFM: A Shared Multimodal Representation of the Moon's Surface (arxiv.org)

Interior interpretability with attention rollout: contraction and propagation profiles in Transformers (arxiv.org)

Indexing: the Beginning and the End (arxiv.org)

Integrated Order Dispatching and Routing for Last-Mile Pickup via Deep Reinforcement Learning (arxiv.org)

IQ-JEPA: A Joint-Embedding Predictive Architecture with a Hermitian Vision Transformer for Sound Speed and Attenuation Estimation from Ultrasound IQ Data (arxiv.org)

Beyond Binary Rooftop Mapping: A Four-Class Deep Learning Framework for Green Roof Potential Assessment from Open Swiss Geospatial Data (arxiv.org)

Cross-Tokenizer On-Policy Distillation via Byte-Prefix Marginalization (arxiv.org)

Evolution-Aware MSA Reasoning for Subsampling via Factor Graphs (arxiv.org)

Synthetic Speech, Real Signal: Paralinguistic Preservation and Cross-Lingual Augmentation via Voice Cloning (arxiv.org)

Efficient Recommendations via Graph Coarsening and Label Propagation (arxiv.org)

Autoregressive EHR Foundation Models with Multimodal Inputs (arxiv.org)

Class-Balanced Softmax: A Bayes Theory-Based Method for Long-Tailed Recognition (arxiv.org)

Optimization of time-consuming experimental conditions using pseudo-experimental data guided by adaptive polynomial regression (arxiv.org)

Latent PDE mapping for efficient physics-informed learning across geometries with limited data (arxiv.org)

Unbiased Open World Regularization for Fair Self-Supervised Learning (arxiv.org)

TriGlue: a Biology-Inspired Generative Model for Generating Molecular Glue-Induced Ternary Complex (arxiv.org)

Pretraining EHR Foundation Models with Patient-Aware Sampling (arxiv.org)