DailyArxiv - AI Research Podcast

Today's papers: - L-TGVN: Leveraging Longitudinal Priors for Personalized Rapid MRI: https://arxiv.org/abs/2606.04419v1 - MapAgent: An Industrial-Grade Agentic Framework for City-scale Lane-level Map Generation: https://arxiv.org/abs/2606.04513v1 - Where Should Knowledge Enter? A Layered Framework for Knowledge Infusion in Multimodal Iterative Generative Mo: https://arxiv.org/abs/2606.06356v1 - Be...

Today's papers: - Inducing Reasoning Primitives from Agent Traces: https://arxiv.org/abs/2606.02994v1 - Self-Refining Agentic Reinforcement Learning for Vision-Conditioned UAV Navigation: https://arxiv.org/abs/2606.03963v1 - MM-BizRAG: Rethinking Multimodal Retrieval-Augmented Generation for General Purpose Enterprise Q&A: https://arxiv.org/abs/2606.04231v1 - ANN Search: Recall What Matters: https...

Today's papers: - Pretraining Language Models on Historical Text: https://arxiv.org/abs/2606.02991v1 - NVIDIA OmniDreams: Real-Time Generative World Model for Closed-Loop Autonomous Vehicle Simulation: https://arxiv.org/abs/2606.03159v1 - LEAP: Supercharging LLMs for Formal Mathematics with Agentic Frameworks: https://arxiv.org/abs/2606.03303v1 - OpenAgenet/OAN: Open Infrastructure for Trusted Age...

Today's papers: - The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence: https://arxiv.org/abs/2605.26494v1 - SeDT: Sentence-Transformer Decision-Transformer Conditioning for Multi-Turn Conversation Reliability: https://arxiv.org/abs/2605.26788v1 - MemFail: Stress-Testing Failure Modes of LLM Memory Systems: https://arxiv.org/abs/2605.26667v1 - Guiding LLM Post-training Da...

Today's papers: - Squeezing Capacity from Multimodal Large Language Models for Subject-driven Generation: https://arxiv.org/abs/2605.26111v1 - ATWL: A Formal Language for Representing, Comparing, and Reusing Visual Analytics Workflows: https://arxiv.org/abs/2605.25489v1 - AI-Assisted Systematization for Evaluating GenAI Systems: https://arxiv.org/abs/2605.26001v1 - PitchBench: Measuring Pitch Hear...

Today's papers: - SkillEvolBench: Benchmarking the Evolution from Episodic Experience to Procedural Skills: https://arxiv.org/abs/2605.24117v1 - Guess the Unified Model: How Much Can We Recover from Generated Images?: https://arxiv.org/abs/2605.25254v1 - Catching MRI outliers: unsupervised detection and localization of MRI artefacts and clinical anomalies using deep learning: https://arxiv.org/abs...

Today's papers: - Transcoders Trace Visual Grounding and Hallucinations in Vision-Language Models: https://arxiv.org/abs/2605.22902v1 - Bernini: Latent Semantic Planning for Video Diffusion: https://arxiv.org/abs/2605.22344v1 - Explainable AI for Data-Driven Design of High-Dimensional Predictive Studies: https://arxiv.org/abs/2605.22243v1 - From TF-IDF to Transformers: A Comparative and Ensemble A...

Today's papers: - Towards a General Intelligence and Interface for Wearable Health Data: https://arxiv.org/abs/2605.22759v1 - What are the Right Symmetries for Formal Theorem Proving?: https://arxiv.org/abs/2605.22257v1 - Not Yet: Humans Outperform LLMs in a Colonel Blotto Tournament: https://arxiv.org/abs/2605.22095v1 - High-speed Networking for Giga-Scale AI Factories: https://arxiv.org/abs/2605...

Today's papers: - EgoBabyVLM: Benchmarking Cross-Modal Learning from Naturalistic Egocentric Video Data - On the limits and opportunities of AI reviewers: Reviewing the reviews of Nature-family papers with 45 expert scientists - ShadeBench: A Benchmark Dataset for Building Shade Simulation in Sustainable Society - Prior Knowledge or Search? A Study of LLM Agents in Hardware-Aware Code Optimization...

Today's papers: - Beyond Inference-Time Search: Reinforcement Learning Synthesizes Reusable Solvers: https://arxiv.org/abs/2605.18374v1 - The Hidden Cost of Contextual Sycophancy: an AI Literacy Intervention in Human-AI Collaboration: https://arxiv.org/abs/2605.18372v1 - CommitDistill: A Lightweight Knowledge-Centric Memory Layer for Software Repositories: https://arxiv.org/abs/2605.18284v1 - Trai...

Today's papers: - Orchard: An Open-Source Agentic Modeling Framework: https://arxiv.org/abs/2605.15040v1 - Cross Modality Image Translation In Medical Imaging Using Generative Frameworks: https://arxiv.org/abs/2605.13686v1 - A Hierarchical Language Model with Predictable Scaling Laws and Provable Benefits of Reasoning: https://arxiv.org/abs/2605.13687v1 - Beyond What to Select: A Plug-and-play Osc...

Today's papers: - Qwen3-VL-Seg: Unlocking Open-World Referring Segmentation with Vision-Language Grounding: https://arxiv.org/abs/2605.07141v1 - Text-to-CAD Evaluation with CADTests: https://arxiv.org/abs/2605.07807v1 - Event-Causal RAG: A Retrieval-Augmented Generation Framework for Long Video Reasoning in Complex Scenarios: https://arxiv.org/abs/2605.06185v1 - Post Reasoning: Improving the Perfo...

Today's papers: - PrefixGuard: From LLM-Agent Traces to Online Failure-Warning Monitors: https://arxiv.org/abs/2605.06455v1 - Quantizing With Randomized Hadamard Transforms: Efficient Heuristic Now Proven: https://arxiv.org/abs/2605.06014v1 - AI Co-Mathematician: Accelerating Mathematicians with Agentic AI: https://arxiv.org/abs/2605.06651v1 - TACT: Mitigating Overthinking and Overacting in Coding...

Today's papers: - The autoPET3 Challenge -- Automated Lesion Segmentation in Whole-Body PET/CT - Multitracer Multicenter Generalization: https://arxiv.org/abs/2605.05775v1 - SCRuB: Social Concept Reasoning under Rubric-Based Evaluation: https://arxiv.org/abs/2605.06444v1 - Delta-Based Neural Architecture Search: LLM Fine-Tuning via Code Diffs: https://arxiv.org/abs/2605.04903v1 - Listwise Policy O...

Today's papers: - LLM-ADAM: A Generalizable LLM Agent Framework for Pre-Print Anomaly Detection in Additive Manufacturing: https://arxiv.org/abs/2605.03328v1 - EvoLM: Self-Evolving Language Models through Co-Evolved Discriminative Rubrics: https://arxiv.org/abs/2605.03871v1 - Exposing LLM Safety Gaps Through Mathematical Encoding:New Attacks and Systematic Analysis: https://arxiv.org/abs/2605.0344...

Today's papers: - From Knowledge to Action: Outcomes of the 2025 Large Language Model (LLM) Hackathon for Applications in Materials Science and Chemistry: https://arxiv.org/abs/2605.03205v1 - Bolek: A Multimodal Language Model for Molecular Reasoning: https://arxiv.org/abs/2605.02745v1 - From Experimental Limits to Physical Insight: A Retrieval-Augmented Multi-Agent Framework for Interpreting Sear...

Today's papers: - Born-Qualified: An Autonomous Framework for Deploying Advanced Energy and Electronic Materials: https://arxiv.org/abs/2605.00639v1 - Towards Multi-Agent Autonomous Reasoning in Hydrodynamics: https://arxiv.org/abs/2605.01102v1 - APIOT: Autonomous Vulnerability Management Across Bare-Metal Industrial OT Networks: https://arxiv.org/abs/2605.02346v1 - NEURON: A Neuro-symbolic System...

Today's papers: - Optimization before Evaluation: Evaluation with Unoptimised Prompts Can be Misleading: https://arxiv.org/abs/2604.27637v1 - Position: agentic AI orchestration should be Bayes-consistent: https://arxiv.org/abs/2605.00742v1 - From Context to Skills: Can Language Models Learn from Context Skillfully?: https://arxiv.org/abs/2604.27660v1 - CastFlow: Learning Role-Specialized Agentic W...

Today's papers: - SecMate: Multi-Agent Adaptive Cybersecurity Troubleshooting with Tri-Context Personalization: https://arxiv.org/abs/2604.26394v1 - Unified 4D World Action Modeling from Video Priors with Asynchronous Denoising: https://arxiv.org/abs/2604.26694v1 - Reliable Answers for Recurring Questions: Boosting Text-to-SQL Accuracy with Template Constrained Decoding: https://arxiv.org/abs/2604...

Today's papers: - From World-Gen to Quest-Line: A Dependency-Driven Prompt Pipeline for Coherent RPG Generation: https://arxiv.org/abs/2604.25482v1 - Agentic Architect: An Agentic AI Framework for Architecture Design Exploration and Optimization: https://arxiv.org/abs/2604.25083v1 - SAFEdit: Does Multi-Agent Decomposition Resolve the Reliability Challenges of Instructed Code Editing?: https://arxi...

Today's papers: - Spectral bandits - Scaling Properties of Continuous Diffusion Spoken Language Models - Nemotron 3 Nano Omni: Efficient and Open Multimodal Intelligence - Rewarding the Scientific Process: Process-Level Reward Modeling for Agentic Data Analysis - MIMIC: A Generative Multimodal Foundation Model for Biomolecules This podcast is from Colin Davis (colin-davis.com) using Claude & Eleve...

Today's papers: - GenMatter: Perceiving Physical Objects with Generative Matter Models: https://arxiv.org/abs/2604.22160v1 - Hard to See, Hard to Label: Generative and Symbolic Acquisition for Subtle Visual Phenomena: https://arxiv.org/abs/2604.22990v1 - LeHome: A Simulation Environment for Deformable Object Manipulation in Household Scenarios: https://arxiv.org/abs/2604.22363v1 - IndustryAssetEQA...

Today's papers: - The First Challenge on Remote Sensing Infrared Image Super-Resolution at NTIRE 2026: Benchmark Results and Method Overview: https://arxiv.org/abs/2604.21312v1 - ChangeQuery: Advancing Remote Sensing Change Analysis for Natural and Human-Induced Disasters from Visual Detection to Semantic Understanding: https://arxiv.org/abs/2604.22333v1 - PermaFrost-Attack: Stealth Pretraining Se...

https://arxiv.org/abs/2604.20329v1 What does it actually mean for a generative visual model to "understand" what it sees? That's the question driving this episode, and it turns out to be harder to answer than it sounds. We start with "Image Generators are Generalist Vision Learners," which introduces VisionBanana, a model built by instruction-tuning NanoBanana Pro on a mix of its original data and...

Today's papers: - Image Generators are Generalist Vision Learners: https://arxiv.org/abs/2604.20329v1 - Centering Ecological Goals in Automated Identification of Individual Animals: https://arxiv.org/abs/2604.20626v1 - BioMiner: A Multi-modal System for Automated Mining of Protein-Ligand Bioactivity Data from Literature: https://arxiv.org/abs/2604.21508v1 - V-tableR1: Process-Supervised Multimodal...