AI Research News Feeds for January 19th, 2026

AI RESEARCH PAPERS & ACADEMIC SOURCES

• SoLA-Vision: Fine-grained Layer-wise Linear Softmax Hybrid Attention
• Democratizing planetary-scale analysis: An ultra-lightweight Earth embedding database for accurate and flexible global land monitoring
• ATATA: One Algorithm to Align Them All
• Bio-inspired fine-tuning for selective transfer learning in image classification
• Image-Text Knowledge Modeling for Unsupervised Multi-Scenario Person Re-Identification
• Language-Agnostic Visual Embeddings for Cross-Script Handwriting Retrieval
• FTDMamba: Frequency-Assisted Temporal Dilation Mamba for Unmanned Aerial Vehicle Video Anomaly Detection
• Efficient On-Board Processing of Oblique UAV Video for Rapid Flood Extent Mapping
• SAMannot: A Memory-Efficient, Local, Open-source Framework for Interactive Video Instance Segmentation based on SAM2
• Context-Aware Semantic Segmentation via Stage-Wise Attention
• Enhancing Vision Language Models with Logic Reasoning for Situational Awareness
• Assessing Building Heat Resilience Using UAV and Street-View Imagery with Coupled Global Context Vision Transformer
• Heterogeneous Uncertainty-Guided Composed Image Retrieval with Fine-Grained Probabilistic Learning
• SUG-Occ: An Explicit Semantics and Uncertainty Guided Sparse Learning Framework for Real-Time 3D Occupancy Prediction
• SME-YOLO: A Real-Time Detector for Tiny Defect Detection on PCB Surfaces
• Generative Scenario Rollouts for End-to-End Autonomous Driving
• ReScene4D: Temporally Consistent Semantic Instance Segmentation of Evolving Indoor 3D Scenes
• UniX: Unifying Autoregression and Diffusion for Chest X-Ray Understanding and Generation
• Differentiating through binarized topology changes: Second-order subpixel-smoothed projection
• KOCOBrain: Kuramoto-Guided Graph Network for Uncovering Structure-Function Coupling in Adolescent Prenatal Drug Exposure
• Convolutions Need Registers Too: HVS-Inspired Dynamic Attention for Video Quality Assessment
• Visual question answering-based image-finding generation for pulmonary nodules on chest CT from structured annotations
• Generation of Chest CT pulmonary Nodule Images by Latent Diffusion Models using the LIDC-IDRI Dataset
• Simple Models, Rich Representations: Visual Decoding from Primate Intracortical Neural Signals
• VidLeaks: Membership Inference Attacks Against Text-to-Video Models
• ProSGNeRF: Progressive Dynamic Neural Scene Graph with Frequency Modulated Foundation Model in Urban Scenes
• Controllable Video Generation: A Survey
• BYOL: Bring Your Own Language Into LLMs
• A Concise Agent is Less Expert: Revealing Side Effects of Using Style Features on Conversational Agents
• EncodeRec: An Embedding Backbone for Recommendation Systems
• DialDefer: A Framework for Detecting and Mitigating LLM Dialogic Deference
• Neural Induction of Finite-State Transducers
• Massively Multilingual Joint Segmentation and Glossing
• ZPD Detector: Data Selection via Capability-Difficulty Alignment for Large Language Models
• Redefining Machine Simultaneous Interpretation: From Incremental Translation to Human-Like Strategies
• NAACL: Noise-AwAre Verbal Confidence Calibration for LLMs in RAG Systems
• From Interpretability to Performance: Optimizing Retrieval Heads for Long-Context Language Models
• Budget-Aware Anytime Reasoning with LLM-Synthesized Preference Data
• Integrity Shield A System for Ethical AI Use & Authorship Transparency in Assessments
• The Growing Gains and Pains of Iterative Web Corpora Crawling: Insights from South Slavic CLASSLA-web 2.0 Corpora
• DOREMI: Optimizing Long Tail Predictions in Document-Level Relation Extraction
• T$^\star$: Progressive Block Scaling for MDM Through Trajectory Aware RL
• MultiCaption: Detecting disinformation using multilingual visual claims
• Language of Thought Shapes Output Diversity in Large Language Models
• One LLM to Train Them All: Multi-Task Learning Framework for Fact-Checking
• Membership Inference on LLMs in the Wild
• F-Actor: Controllable Conversational Behaviour in Full-Duplex Models
• Idea First, Code Later: Disentangling Problem Solving from Code Generation in Evaluating LLMs for Competitive Programming
• Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Models
• Reward Modeling for Scientific Writing Evaluation
• The unreasonable effectiveness of pattern matching
• Predict the Retrieval! Test time adaptation for Retrieval Augmented Generation
• CTest-Metric: A Unified Framework to Assess Clinical Validity of Metrics for CT Report Generation
• How Long Is a Piece of String? A Brief Empirical Analysis of Tokenizers
• AJAR: Adaptive Jailbreak Architecture for Red-teaming
• SonicBench: Dissecting the Physical Perception Bottleneck in Large Audio Language Models
• FlashLabs Chroma 1.0: A Real-Time End-to-End Spoken Dialogue Model with Personalized Voice Cloning
• Isotropy-Optimized Contrastive Learning for Semantic Course Recommendation
• Future Optical Flow Prediction Improves Robot Control & Video Generation
• ICONIC-444: A 3.1-Million-Image Dataset for OOD Detection Research
• A Unified 3D Object Perception Framework for Real-Time Outside-In Multi-Camera Systems
• One Model, Many Behaviors: Training-Induced Effects on Out-of-Distribution Detection
• Effects of Different Attention Mechanisms Applied on 3D Models in Video Classification
• FrankenMotion: Part-level Human Motion Generation and Composition
• Classification of Chest XRay Diseases through image processing and analysis techniques
• MMedExpert-R1: Strengthening Multimodal Medical Reasoning via Domain-Specific Adaptation and Clinical Guideline Reinforcement
• M3DDM+: An improved video outpainting by a modified masking strategy
• PhysRVG: Physics-Aware Unified Reinforcement Learning for Video Generative Models
• CoDance: An Unbind-Rebind Paradigm for Robust Multi-Subject Animation
• Graph Smoothing for Enhanced Local Geometry Learning in Point Cloud Analysis
• Operator learning on domain boundary through combining fundamental solution-based artificial data and boundary integral techniques
• Latent Dynamics Graph Convolutional Networks for model order reduction of parameterized time-dependent PDEs
• Sample-Near-Optimal Agnostic Boosting with Improved Running Time
• Metabolomic Biomarker Discovery for ADHD Diagnosis Using Interpretable Machine Learning
• FORESTLLM: Large Language Models Make Random Forest Great on Few-shot Tabular Learning
• Unlocking the Potentials of Retrieval-Augmented Generation for Diffusion Language Models
• Offline Reinforcement-Learning-Based Power Control for Application-Agnostic Energy Efficiency
• Latent Space Inference via Paired Autoencoders
• Factored Value Functions for Graph-Based Multi-Agent Reinforcement Learning
• Forcing and Diagnosing Failure Modes of Fourier Neural Operators Across Diverse PDE Families
• Inter-patient ECG Arrhythmia Classification with LGNs and LUTNs
• When Are Two Scores Better Than One? Investigating Ensembles of Diffusion Models
• Low-Rank Key Value Attention
• Extractive summarization on a CMOS Ising machine
• QUPID: A Partitioned Quantum Neural Network for Anomaly Detection in Smart Grid
• SSC-UNet: UNet with Self-Supervised Contrastive Learning for Phonocardiography Noise Reduction
• UBiGTLoc: A Unified BiLSTM-Graph Transformer Localization Framework for IoT Sensor Networks
• Sensor Placement for Urban Traffic Interpolation: A Data-Driven Evaluation to Inform Policy
• Mass Distribution versus Density Distribution in the Context of Clustering
• Physically constrained unfolded multi-dimensional OMP for large MIMO systems
• LLMs for Game Theory: Entropy-Guided In-Context Learning and Adaptive CoT Reasoning
• Reasoning Models Generate Societies of Thought
• Learning collision operators from plasma phase space data using differentiable simulators
• A PAC-Bayesian Analysis of Channel-Induced Degradation in Edge Inference
• Depression Detection Based on Electroencephalography Using a Hybrid Deep Neural Network CNN-GRU and MRMR Feature Selection
• Memorize Early, Then Query: Inlier-Memorization-Guided Active Outlier Detection
• Exact Constraint Enforcement in Physics-Informed Extreme Learning Machines using Null-Space Projection Framework
• CoG: Controllable Graph Reasoning via Relational Blueprints and Failure-Aware Refinement over Knowledge Graphs
• Split-and-Conquer: Distributed Factor Modeling for High-Dimensional Matrix-Variate Time Series
• KANHedge: Efficient Hedging of High-Dimensional Options Using Kolmogorov-Arnold Network-Based BSDE Solver
• Comprehensive Robust Dynamic Mode Decomposition from Mode Extraction to Dimensional Reduction
• Model-free policy gradient for discrete-time mean-field control
• How DDAIR you? Disambiguated Data Augmentation for Intent Recognition
• Reasoning in Trees: Improving Retrieval-Augmented Generation for Multi-Hop Question Answering
• Effects of Introducing Synaptic Scaling on Spiking Neural Network Learning
• Scalable Music Cover Retrieval Using Lyrics-Aligned Audio Embeddings
• Information Theoretic Perspective on Representation Learning
• Beer-Lambert Autoencoder for Unsupervised Stain Representation Learning and Deconvolution in Multi-immunohistochemical Brightfield Histology Images
• New Adaptive Mechanism for Large Neighborhood Search using Dual Actor-Critic
• Zero-Shot Detection of Elastic Transient Morphology Across Physical Systems
• Statistical Robustness of Interval CVaR Based Regression Models under Perturbation and Contamination
• PubMed-OCR: PMC Open Access OCR Annotations
• Near-Optimal Decentralized Stochastic Nonconvex Optimization with Heavy-Tailed Noise
• IMS: Intelligent Hardware Monitoring System for Secure SoCs
• Learning Semantic-Geometric Task Graph-Representations from Human Demonstrations
• A Probabilistic Approach to Trajectory-Based Optimal Experimental Design
• On the Probability of First Success in Differential Evolution: Hazard Identities and Tail Bounds
• ShapeR: Robust Conditional 3D Shape Generation from Casual Captures
• ThinkEval: Practical Evaluation of Knowledge Leakage in LLM Editing using Thought-based Knowledge Graphs
• UCB-type Algorithm for Budget-Constrained Expert Learning
• Detecting Toxic Flow
• High-Dimensional Tail Index Regression
• A Natural Primal-Dual Hybrid Gradient Method for Adversarial Neural Network Training on Solving Partial Differential Equations
• Conditional Distribution Compression via the Kernel Conditional Mean Embedding
• Feature Propagation on Knowledge Graphs using Cellular Sheaves
• Theorem Prover as a Judge for Synthetic Data Generation
• Utilizing Class Separation Distance for the Evaluation of Corruption Robustness of Machine Learning Classifiers
• A Simple Unified Uncertainty-Guided Framework for Offline-to-Online Reinforcement Learning
• Balanced Edge Pruning for Graph Anomaly Detection with Noisy Labels
• Policy alone is probably not the solution: A large-scale experiment on how developers struggle to design meaningful end-user explanations
• Generalizable Domain Adaptation for Sim-and-Real Policy Co-Training
• Vendor-Aware Industrial Agents: RAG-Enhanced LLMs for Secure On-Premise PLC Code Generation
• Analytic Bijections for Smooth and Interpretable Normalizing Flows
• Towards Tensor Network Models for Low-Latency Jet Tagging on FPGAs
• Mugi: Value Level Parallelism For Efficient LLMs
• AI-Guided Human-In-the-Loop Inverse Design of High Performance Engineering Structures
• Beyond Accuracy: A Stability-Aware Metric for Multi-Horizon Forecasting
• Unit-Consistent (UC) Adjoint for GSD and Backprop in Deep Learning Applications
• Action Shapley: A Training Data Selection Metric for World Model in Reinforcement Learning
• Realistic Curriculum Reinforcement Learning for Autonomous and Sustainable Marine Vessel Navigation
• FAConvLSTM: Factorized-Attention ConvLSTM for Efficient Feature Extraction in Multivariate Climate Data
• HOSL: Hybrid-Order Split Learning for Memory-Constrained Edge Training
• Multivariate LSTM-Based Forecasting for Renewable Energy: Enhancing Climate Change Mitigation
• Transient learning dynamics drive escape from sharp valleys in Stochastic Gradient Descent
• Toward Adaptive Grid Resilience: A Gradient-Free Meta-RL Framework for Critical Load Restoration
• Reasoning Distillation for Lightweight Automated Program Repair
• Constant Metric Scaling in Riemannian Computation
• Backdoor Attacks on Multi-modal Contrastive Learning
• Matching High-Dimensional Geometric Quantiles for Test-Time Adaptation of Transformers and Convolutional Networks Alike
• AVP-Pro: An Adaptive Multi-Modal Fusion and Contrastive Learning Approach for Comprehensive Two-Stage Antiviral Peptide Identification
• Self-Augmented Mixture-of-Experts for QoS Prediction
• OpFML: Pipeline for ML-based Operational Forecasting
• Spurious Rewards Paradox: Mechanistically Understanding How RLVR Activates Memorization Shortcuts in LLMs
• Soft Bayesian Context Tree Models for Real-Valued Time Series
• Differentially Private Subspace Fine-Tuning for Large Language Models
• Optimized Algorithms for Text Clustering with LLM-Generated Constraints
• Shape-morphing programming of soft materials on complex geometries via neural operator
• FSL-BDP: Federated Survival Learning with Bayesian Differential Privacy for Credit Risk Modeling
• Assesing the Viability of Unsupervised Learning with Autoencoders for Predictive Maintenance in Helicopter Engines
• Theoretically and Practically Efficient Resistance Distance Computation on Large Graphs
• GMM-COMET: Continual Source-Free Universal Domain Adaptation via a Mean Teacher and Gaussian Mixture Model-Based Pseudo-Labeling
• LSTM VS. Feed-Forward Autoencoders for Unsupervised Fault Detection in Hydraulic Pumps
• TimeMar: Multi-Scale Autoregressive Modeling for Unconditional Time Series Generation
• Exploring LLM Features in Predictive Process Monitoring for Small-Scale Event-Logs
• Health Facility Location in Ethiopia: Leveraging LLMs to Integrate Expert Knowledge into Algorithmic Planning
• BoxMind: Closed-loop AI strategy optimization for elite boxing validated in the 2024 Olympics
• Generative AI Purpose-built for Social and Mental Health: A Real-World Pilot
• EvidFuse: Writing-Time Evidence Learning for Consistent Text-Chart Data Reporting
• DSA-Tokenizer: Disentangled Semantic-Acoustic Tokenization via Flow Matching-based Hierarchical Fusion
• Millimeter-Wave Gesture Recognition in ISAC: Does Reducing Sensing Airtime Hamper Accuracy?
• Neuro-Symbolic Activation Discovery: Transferring Mathematical Structures from Physics to Ecology for Parameter-Efficient Neural Networks
• Line-based Event Preprocessing: Towards Low-Energy Neuromorphic Computer Vision
• AnyECG: Evolved ECG Foundation Model for Holistic Health Profiling
• Unifying Speech Recognition, Synthesis and Conversion with Autoregressive Transformers
• LogicLens: Leveraging Semantic Code Graph to explore Multi Repository large systems
• Unified Optimization of Source Weights and Transfer Quantities in Multi-Source Transfer Learning: An Asymptotic Framework
• Digital Metabolism: Decoupling Logic from Facts via Regenerative Unlearning -- Towards a Pure Neural Logic Core
• Towards Reliable ML Feature Engineering via Planning in Constrained-Topology of LLM Agents
• Approximately Optimal Global Planning for Contact-Rich SE(2) Manipulation on a Graph of Reachable Sets
• Can Vision-Language Models Understand Construction Workers? An Exploratory Study
• Medical SAM3: A Foundation Model for Universal Prompt-Driven Medical Image Segmentation
• Self-learned representation-guided latent diffusion model for breast cancer classification in deep ultraviolet whole surface images
• RobuMTL: Enhancing Multi-Task Learning Robustness Against Weather Conditions
• Selecting Language Models for Social Science: Start Small, Start Open, and Validate
• Sparse Data Tree Canopy Segmentation: Fine-Tuning Leading Pretrained Models on Only 150 Images
• PatientVLM Meets DocVLM: Pre-Consultation Dialogue Between Vision-Language Models for Efficient Diagnosis
• Multi-Stage Patient Role-Playing Framework for Realistic Clinical Interactions
• Beyond Max Tokens: Stealthy Resource Amplification via Tool Calling Chains in LLM Agents
• Steering Language Models Before They Speak: Logit-Level Interventions
• When Personalization Misleads: Understanding and Mitigating Hallucinations in Personalized LLMs
• Contextual Distributionally Robust Optimization with Causal and Continuous Structure: An Interpretable and Tractable Approach
• Finding the Translation Switch: Discovering and Exploiting the Task-Initiation Features in LLMs
• Combating Spurious Correlations in Graph Interpretability via Self-Reflection
• IDDR-NGP: Incorporating Detectors for Distractor Removal with Instant Neural Radiance Field
• Your One-Stop Solution for AI-Generated Video Detection
• Spectral Characterization and Mitigation of Sequential Knowledge Editing Collapse
• Predicting Biased Human Decision-Making with Large Language Models in Conversational Settings
• H-AIM: Orchestrating LLMs, PDDL, and Behavior Trees for Hierarchical Multi-Robot Planning
• Fairness in Healthcare Processes: A Quantitative Analysis of Decision Making in Triage
• Bridging Cognitive Neuroscience and Graph Intelligence: Hippocampus-Inspired Multi-View Hypergraph Learning for Web Finance Fraud
• A3D: Adaptive Affordance Assembly with Dual-Arm Manipulation
• ABC-Bench: Benchmarking Agentic Backend Coding in Real-World Development
• Visual Marker Search for Autonomous Drone Landing in Diverse Urban Environments
• Efficient Multilingual Name Type Classification Using Convolutional Networks
• Vision-as-Inverse-Graphics Agent via Interleaved Multimodal Reasoning
• Learn Before Represent: Bridging Generative and Contrastive Learning for Domain-Specific LLM Embeddings
• Context-aware Graph Causality Inference for Few-Shot Molecular Property Prediction
• Learning Quadrupedal Locomotion for a Heavy Hydraulic Robot Using an Actuator Model
• Deep GraphRAG: A Balanced Approach to Hierarchical Retrieval and Adaptive Integration
• Cross-Modal Attention Network with Dual Graph Learning in Multimodal Recommendation
• Clustering High-dimensional Data: Balancing Abstraction and Representation Tutorial at AAAI 2026
• Artificial Intelligence and the US Economy: An Accounting Perspective on Investment and Production
• SD-RAG: A Prompt-Injection-Resilient Framework for Selective Disclosure in Retrieval-Augmented Generation
• FAQ: Mitigating Quantization Error via Regenerating Calibration Data with Family-Aware Quantization
• Epistemic Control and the Normativity of Machine Learning-Based Science
• LoRA as Oracle
• SDFLoRA: Selective Dual-Module LoRA for Federated Fine-tuning with Heterogeneous Clients
• FactCorrector: A Graph-Inspired Approach to Long-Form Factuality Correction of Large Language Models
• Knowledge is Not Enough: Injecting RL Skills for Continual Adaptation
• X-Distill: Cross-Architecture Vision Distillation for Visuomotor Learning
• From SERPs to Sound: How Search Engine Result Pages and AI-generated Podcasts Interact to Influence User Attitudes on Controversial Topics
• How Much Would a Clinician Edit This Draft? Evaluating LLM Alignment for Patient Message Response Drafting
• FEATHer: Fourier-Efficient Adaptive Temporal Hierarchy Forecaster for Time-Series Forecasting
• Think-Clip-Sample: Slow-Fast Frame Selection for Video Understanding
• Institutional AI: Governing LLM Collusion in Multi-Agent Cournot Markets via Public Governance Graphs
• Evaluating LLM Behavior in Hiring: Implicit Weights, Fairness Across Groups, and Alignment with Human Preferences
• Wetland mapping from sparse annotations with satellite image time series and temporal-aware segment anything model
• Topology-Guaranteed Image Segmentation: Enforcing Connectivity, Genus, and Width Constraints
• The Great March 100: 100 Detail-oriented Tasks for Evaluating Embodied AI Agents
• Relational Linearity is a Predictor of Hallucinations
• GenDA: Generative Data Assimilation on Complex Urban Areas via Classifier-Free Diffusion Guidance
• Hierarchical Orthogonal Residual Spread for Precise Massive Editing in Large Language Models
• Map2Thought: Explicit 3D Spatial Reasoning via Metric Cognitive Maps
• PRISM-CAFO: Prior-conditioned Remote-sensing Infrastructure Segmentation and Mapping for CAFOs
• Interactive Narrative Analytics: Bridging Computational Narrative Extraction and Human Sensemaking
• MHA2MLA-VLM: Enabling DeepSeek's Economical Multi-Head Latent Attention across Vision-Language Models
• The Poisoned Apple Effect: Strategic Manipulation of Mediated Markets via Technology Expansion of AI Agents
• MetaboNet: The Largest Publicly Available Consolidated Dataset for Type 1 Diabetes Management
• Building Production-Ready Probes For Gemini
• Do explanations generalize across large reasoning models?
• Japanese AI Agent System on Human Papillomavirus Vaccination: System Design
• Do You Trust Me? Cognitive-Affective Signatures of Trustworthiness in Large Language Models
• Building AI Agents to Improve Job Referral Requests to Strangers
• ORBITFLOW: SLO-Aware Long-Context LLM Serving with Fine-Grained KV Cache Reconfiguration
• CTHA: Constrained Temporal Hierarchical Architecture for Stable Multi-Agent LLM Systems
• Explore with Long-term Memory: A Benchmark and Multimodal LLM-based Reinforcement Learning Framework for Embodied Exploration
• Optimisation of complex product innovation processes based on trend models with three-valued logic
• ARC Prize 2025: Technical Report
• What Matters in Data Curation for Multimodal Reasoning? Insights from the DCVLR Challenge
• AdaMARP: An Adaptive Multi-Agent Interaction Framework for General Immersive Role-Playing
• Efficient Protein Optimization via Structure-aware Hamiltonian Dynamics
• BAPO: Boundary-Aware Policy Optimization for Reliable Agentic Search
• AgencyBench: Benchmarking the Frontiers of Autonomous Agents in 1M-Token Real-World Contexts
• MiCA: A Mobility-Informed Causal Adapter for Lightweight Epidemic Forecasting
• ReCreate: Reasoning and Creating Domain Agents Driven by Experience
• Do We Always Need Query-Level Workflows? Rethinking Agentic Workflow Generation for Multi-Agent Systems
• TANDEM: Temporal-Aware Neural Detection for Multimodal Hate Speech
• Policy-Based Deep Reinforcement Learning Hyperheuristics for Job-Shop Scheduling Problems
• Beyond Model Scaling: Test-Time Intervention for Efficient Deep Reasoning
• XChoice: Explainable Evaluation of AI-Human Alignment in LLM-based Constrained Choice Decision Making
• AstroReason-Bench: Evaluating Unified Agentic Planning across Heterogeneous Space Planning Problems
• Hyperparameter Optimization of Constraint Programming Solvers