{ "schemaVersion": "rnn.research.v1", "generatedAt": "2026-05-26T00:07:16.321Z", "count": 97, "papers": [ { "id": "arxiv-query-cat-cs-ro-468314c4:WqTw4LNkwcn_uVL2A5", "title": "Good Token Hunting: A Hitchhiker's Guide to Token Selection for Visual Geometry Transformers", "authors": [ "Shuhong Zheng" ], "abstract": "Visual geometry transformers have become powerful architectures for multi-view 3D reconstruction, enabling joint prediction of multiple 3D attributes in a feed-forward manner.", "arxivId": "2605.23892v1", "url": "https://arxiv.org/abs/2605.23892v1", "pdfUrl": "https://arxiv.org/pdf/2605.23892", "categories": [ "robotics-research", "Robotics", "cs.CV", "cs.AI", "cs.GR" ], "publishedAt": "2026-05-22T17:55:13.000Z", "summary": "Visual geometry transformers have become powerful architectures for multi-view 3D reconstruction, enabling joint prediction of multiple 3D attributes in a feed-forward manner.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-ro-468314c4:-SZgnsf8R3-mj-rJ-D", "title": "Robotic Strawberry Harvesting with Robust Vision and Deep Reinforcement Learning based Sim-to-Real Control", "authors": [ "Al Bashir" ], "abstract": "This study presents a closed-loop robotic strawberry harvesting system that combines a robust vision module, simulation-trained deep reinforcement learning (DRL) control, and ROS-based realrobot execution.", "arxivId": "2605.23863v1", "url": "https://arxiv.org/abs/2605.23863v1", "pdfUrl": "https://arxiv.org/pdf/2605.23863", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-22T17:21:06.000Z", "summary": "This study presents a closed-loop robotic strawberry harvesting system that combines a robust vision module, simulation-trained deep reinforcement learning (DRL) control, and ROS-based realrobot execution.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-ro-468314c4:veIIHZa84EBLqNighx", "title": "Point Tracking Improves World Action Models", "authors": [ "Jiarui Guan" ], "abstract": "Robot policy learning benefits from world-action models that capture environment dynamics, but pixel-level prediction entangles dynamics with nuisance factors such as lighting and texture, making learned representations vulnerable to task-irrelevant visual...", "arxivId": "2605.23856v1", "url": "https://arxiv.org/abs/2605.23856v1", "pdfUrl": "https://arxiv.org/pdf/2605.23856", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-22T17:08:37.000Z", "summary": "Robot policy learning benefits from world-action models that capture environment dynamics, but pixel-level prediction entangles dynamics with nuisance factors such as lighting and texture, making learned representations vulnerable to task-irrelevant visual...", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-ro-468314c4:3D4teM2V2n3enF_fMy", "title": "Instrumentation for Imitation Learning: Enhancing Training Datasets for Clothes Hanger Insertion", "authors": [ "Remko Proesmans" ], "abstract": "Large behaviour models have transformed the field of robotic manipulation, but prohibitive data requirements have thus far prevented a revolution similar to vision language models.", "arxivId": "2605.23847v1", "url": "https://arxiv.org/abs/2605.23847v1", "pdfUrl": "https://arxiv.org/pdf/2605.23847", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-22T16:59:55.000Z", "summary": "Large behaviour models have transformed the field of robotic manipulation, but prohibitive data requirements have thus far prevented a revolution similar to vision language models.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-ro-468314c4:AAcsxjb5ELvKX1S8C9", "title": "SFG-ROS: A Resource-Aware Framework for Dense Multi-Agent Perception", "authors": [ "Constantin Blessing" ], "abstract": "Deploying heterogeneous multi-agent robot fleets for collaborative perception requires robust data exchange and scalable software architectures.", "arxivId": "2605.23832v1", "url": "https://arxiv.org/abs/2605.23832v1", "pdfUrl": "https://arxiv.org/pdf/2605.23832", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-22T16:41:12.000Z", "summary": "Deploying heterogeneous multi-agent robot fleets for collaborative perception requires robust data exchange and scalable software architectures.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-imitation-learning-robot-cca50e90:4QZW2kfRtbT_bGVtOG", "title": "Direct Dynamic Retargeting for Humanoid Imitation Learning from Videos", "authors": [ "Constant Roux" ], "abstract": "Imitation Learning from monocular video demonstrations provides a scalable approach for teaching complex skills to humanoid robots.", "arxivId": "2605.23762v1", "url": "https://arxiv.org/abs/2605.23762v1", "pdfUrl": "https://arxiv.org/pdf/2605.23762", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-22T15:33:40.000Z", "summary": "Imitation Learning from monocular video demonstrations provides a scalable approach for teaching complex skills to humanoid robots.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-foundation-model-0634b051:ZE8ycix-KP3Tgcwty9", "title": "Any2Any: Efficient Cross-Embodiment Transfer for Humanoid Whole-Body Tracking", "authors": [ "Ming Yang" ], "abstract": "Whole-body tracking (WBT) models have become a key foundation for humanoid robots, enabling them to imitate diverse motions with high fidelity.", "arxivId": "2605.23733v1", "url": "https://arxiv.org/abs/2605.23733v1", "pdfUrl": "https://arxiv.org/pdf/2605.23733", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.AI" ], "publishedAt": "2026-05-22T15:10:42.000Z", "summary": "Whole-body tracking (WBT) models have become a key foundation for humanoid robots, enabling them to imitate diverse motions with high fidelity.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-manipulation-4d4e16c1:QmI7DmwYarmTDhdDKk", "title": "How Many Training Samples Are Needed for the Inverse Kinematics Solutions by Artificial Neural Networks", "authors": [ "Dong-Won Lim" ], "abstract": "Inverse Kinematics (IK) plays a critical role in robotic motion planning and control.", "arxivId": "2605.23583v1", "url": "https://arxiv.org/abs/2605.23583v1", "pdfUrl": "https://arxiv.org/pdf/2605.23583", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.LG" ], "publishedAt": "2026-05-22T12:54:08.000Z", "summary": "Inverse Kinematics (IK) plays a critical role in robotic motion planning and control.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-manipulation-4d4e16c1:HW7zi99XAmiEuP4Sgl", "title": "ComPose: When to Trust Hands for Object Pose Tracking", "authors": [ "Jisu Shin" ], "abstract": "Reconstructing the motion of objects from videos is a key component for embodied AI and robot manipulation.", "arxivId": "2605.23523v1", "url": "https://arxiv.org/abs/2605.23523v1", "pdfUrl": "https://arxiv.org/pdf/2605.23523", "categories": [ "robotics-research", "Robotics", "cs.CV" ], "publishedAt": "2026-05-22T11:39:49.000Z", "summary": "Reconstructing the motion of objects from videos is a key component for embodied AI and robot manipulation.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-mobile-manipulation-ec740706:4miZnMIQXVdAMbzbn3", "title": "Sparse Compositional Flow Matching by geometric assembly from motion primitives", "authors": [ "Yan Tang" ], "abstract": "Embodied trajectories, such as the executable motion sequences of robotic manipulators, underwater vehicles, and mobile robots, are a fundamental output of embodied AI.", "arxivId": "2605.23341v1", "url": "https://arxiv.org/abs/2605.23341v1", "pdfUrl": "https://arxiv.org/pdf/2605.23341", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.AI" ], "publishedAt": "2026-05-22T07:55:48.000Z", "summary": "Embodied trajectories, such as the executable motion sequences of robotic manipulators, underwater vehicles, and mobile robots, are a fundamental output of embodied AI.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-humanoid-robot-bebcca07:87LdhGIIlI4NP9UGv8", "title": "Signal Temporal Logic Motion Planning via Graphs of Convex Sets", "authors": [ "Yu Chen" ], "abstract": "This paper investigates continuous-time motion planning under Signal Temporal Logic (STL) specifications.", "arxivId": "2605.23240v1", "url": "https://arxiv.org/abs/2605.23240v1", "pdfUrl": "https://arxiv.org/pdf/2605.23240", "categories": [ "robotics-research", "Robotics", "cs.RO", "eess.SY" ], "publishedAt": "2026-05-22T05:19:43.000Z", "summary": "This paper investigates continuous-time motion planning under Signal Temporal Logic (STL) specifications.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-embodied-ai-00b1f9b2:DhO7CJ83N8zrw47LYu", "title": "AutoResearch AI: Towards AI Research Automation for Scientific Discovery", "authors": [ "Guiyao Tie" ], "abstract": "Scientific research is being reshaped by AI systems that move beyond isolated assistance toward longer-horizon workflows spanning literature grounding, hypothesis generation, experimentation, validation, reporting, and revision.", "arxivId": "2605.23204v1", "url": "https://arxiv.org/abs/2605.23204v1", "pdfUrl": "https://arxiv.org/pdf/2605.23204", "categories": [ "robotics-research", "Robotics", "cs.AI" ], "publishedAt": "2026-05-22T03:40:30.000Z", "summary": "Scientific research is being reshaped by AI systems that move beyond isolated assistance toward longer-horizon workflows spanning literature grounding, hypothesis generation, experimentation, validation, reporting, and revision.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-embodied-ai-00b1f9b2:DlIh8rtl7FopPZhxN2", "title": "IntentionNav: A Benchmark for Intent-Driven Object Navigation from Implicit Human Instruction", "authors": [ "Lin Qian" ], "abstract": "Existing object navigation benchmarks usually tell an embodied agent which object category to find, such as microwave or chair.", "arxivId": "2605.23187v1", "url": "https://arxiv.org/abs/2605.23187v1", "pdfUrl": "https://arxiv.org/pdf/2605.23187", "categories": [ "robotics-research", "Robotics", "cs.CV", "cs.RO" ], "publishedAt": "2026-05-22T03:09:55.000Z", "summary": "Existing object navigation benchmarks usually tell an embodied agent which object category to find, such as microwave or chair.", "relevanceScore": 0.75, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-model-context-protocol-1cc7e65f:1iAM4z4LvnLLUemXOj", "title": "World Machine: Towards Generative World Modeling for Time-Series", "authors": [ "Elton Cardoso do Nascimento" ], "abstract": "World models represent a paradigm shift in generative AI, pursuing predictive understanding and controllable simulation of environments in a structured and generalizable way.", "arxivId": "2605.23025v1", "url": "https://arxiv.org/abs/2605.23025v1", "pdfUrl": "https://arxiv.org/pdf/2605.23025", "categories": [ "robotics-research", "Robotics", "cs.LG" ], "publishedAt": "2026-05-21T20:48:51.000Z", "summary": "World models represent a paradigm shift in generative AI, pursuing predictive understanding and controllable simulation of environments in a structured and generalizable way.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-model-context-protocol-1cc7e65f:cYIO9BznujVxoRzlnH", "title": "A Reproducible Universal Dependencies-Style Pipeline for Katharevousa Greek Parliamentary Text", "authors": [ "George Mikros" ], "abstract": "Katharevousa Greek remains poorly served by contemporary NLP pipelines despite its importance for legal, administrative, and parliamentary archives.", "arxivId": "2605.22978v1", "url": "https://arxiv.org/abs/2605.22978v1", "pdfUrl": "https://arxiv.org/pdf/2605.22978", "categories": [ "robotics-research", "Robotics", "cs.CL" ], "publishedAt": "2026-05-21T19:16:20.000Z", "summary": "Katharevousa Greek remains poorly served by contemporary NLP pipelines despite its importance for legal, administrative, and parliamentary archives.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-foundation-model-0634b051:6XLq7VOKYO_vsPv8HR", "title": "Hollow Needle Puncture Mechanics for Biopsy Sampling", "authors": [ "Yiting Wu" ], "abstract": "Biopsy sampling relies on hollow needles that puncture soft tissues by propagating and opening a cylindrical crack, yet the mechanics governing this coring process remain only partially understood.", "arxivId": "2605.22790v1", "url": "https://arxiv.org/abs/2605.22790v1", "pdfUrl": "https://arxiv.org/pdf/2605.22790", "categories": [ "robotics-research", "Robotics", "cond-mat.soft", "cond-mat.mtrl-sci" ], "publishedAt": "2026-05-21T17:44:21.000Z", "summary": "Biopsy sampling relies on hollow needles that puncture soft tissues by propagating and opening a cylindrical crack, yet the mechanics governing this coring process remain only partially understood.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-mobile-manipulation-ec740706:-YSpATsVsVYZ3b7KMD", "title": "Transport Enhancement and In Situ Control of Electronic Correlation via Photoinduced Modulation Doping of van der Waals Heterostructures", "authors": [ "Collin R. Sanborn" ], "abstract": "Modulation doping, a well-established technique for traditional semiconductor heterostructures, is a promising approach for tailoring carrier concentration in 2D materials devices.", "arxivId": "2605.22452v1", "url": "https://arxiv.org/abs/2605.22452v1", "pdfUrl": "https://arxiv.org/pdf/2605.22452", "categories": [ "robotics-research", "Robotics", "cond-mat.mes-hall" ], "publishedAt": "2026-05-21T13:17:57.000Z", "summary": "Modulation doping, a well-established technique for traditional semiconductor heterostructures, is a promising approach for tailoring carrier concentration in 2D materials devices.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-teleoperation-robot-9f4f2b08:utZ7gtZ0xI26igUh04", "title": "Remote Teleoperation of Endovascular Intervention Robots: A Systematic Review", "authors": [ "Xingyu Chen" ], "abstract": "Remote robotic-assisted endovascular intervention offers a promising approach to reduce clinician radiation exposure and physical strain, while extending specialized vascular care to geographically distant regions.", "arxivId": "2605.22889v1", "url": "https://arxiv.org/abs/2605.22889v1", "pdfUrl": "https://arxiv.org/pdf/2605.22889", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-21T10:38:11.000Z", "summary": "Remote robotic-assisted endovascular intervention offers a promising approach to reduce clinician radiation exposure and physical strain, while extending specialized vascular care to geographically distant regions.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-foundation-model-0634b051:bmhnKwzdXOUmR9qQON", "title": "Imagine2Real: Towards Zero-shot Humanoid-Object Interaction via Video Generative Priors", "authors": [ "Jiahe Chen" ], "abstract": "Whole-body Humanoid-Object Interaction (HOI) is bottlenecked by the scarcity of high-fidelity 3D data.", "arxivId": "2605.22272v2", "url": "https://arxiv.org/abs/2605.22272v2", "pdfUrl": "https://arxiv.org/pdf/2605.22272", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.CV" ], "publishedAt": "2026-05-21T10:15:39.000Z", "summary": "Whole-body Humanoid-Object Interaction (HOI) is bottlenecked by the scarcity of high-fidelity 3D data.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-humanoid-robot-bebcca07:pUfxwk4uV8EbjydkfZ", "title": "Learning to Evolve: Multi-modal Interactive Fields for Robust Humanoid Navigation in Dynamic Environments", "authors": [ "Peifeng Jiang" ], "abstract": "Safe manipulation-oriented navigation for humanoid robots requires scene memory that remains reliable under locomotion-induced perceptual distortion, environmental changes, and interaction-level geometric safety constraints.", "arxivId": "2605.21935v1", "url": "https://arxiv.org/abs/2605.21935v1", "pdfUrl": "https://arxiv.org/pdf/2605.21935", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-21T03:11:43.000Z", "summary": "Safe manipulation-oriented navigation for humanoid robots requires scene memory that remains reliable under locomotion-induced perceptual distortion, environmental changes, and interaction-level geometric safety constraints.", "relevanceScore": 0.93, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "Robotics safety", "slug": "robotics-safety", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-mobile-manipulation-ec740706:FzWteqi6Er81daG8jm", "title": "Motion Design for Grasp-Based Dynamic Locomotion in Microgravity", "authors": [ "Chaerim Moon" ], "abstract": "Locomotion in microgravity often relies on sparsely and irregularly arranged anchors, motivating grasp-based mobility with multiple limbs.", "arxivId": "2605.21704v1", "url": "https://arxiv.org/abs/2605.21704v1", "pdfUrl": "https://arxiv.org/pdf/2605.21704", "categories": [ "robotics-research", "Robotics", "cs.RO", "eess.SY" ], "publishedAt": "2026-05-20T20:06:01.000Z", "summary": "Locomotion in microgravity often relies on sparsely and irregularly arranged anchors, motivating grasp-based mobility with multiple limbs.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-humanoid-robot-bebcca07:4I9THQtEn-lehlcXEi", "title": "To Select or not to Select, that is the Question: Distilling Robot Skill Prediction into a Small Ensemble", "authors": [ "Haechan Mark Bong" ], "abstract": "As robot fleets become more heterogeneous, including humanoids, rovers, quadrupeds, and drones, selecting the right robot for a task becomes a core systems problem.", "arxivId": "2605.21242v1", "url": "https://arxiv.org/abs/2605.21242v1", "pdfUrl": "https://arxiv.org/pdf/2605.21242", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-20T14:32:38.000Z", "summary": "As robot fleets become more heterogeneous, including humanoids, rovers, quadrupeds, and drones, selecting the right robot for a task becomes a core systems problem.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-mobile-manipulation-ec740706:pjCb1GEobRJM3QTIrE", "title": "WiXus: A Wheeled-Legged Robot with Wire-Driven Environmental Utilizing to Integrate Mobility and Manipulation", "authors": [ "Shintaro Inoue" ], "abstract": "Wheeled-legged robots, which have wheels at their feet and achieve high mobility by coordinating wheel drive and leg drive, have been developed.", "arxivId": "2605.20932v1", "url": "https://arxiv.org/abs/2605.20932v1", "pdfUrl": "https://arxiv.org/pdf/2605.20932", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-20T09:19:48.000Z", "summary": "Wheeled-legged robots, which have wheels at their feet and achieve high mobility by coordinating wheel drive and leg drive, have been developed.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-teleoperation-robot-9f4f2b08:AX-KkZGw9qq-i9zkTW", "title": "SUGAR: A Scalable Human-Video-Driven Generalizable Humanoid Loco-Manipulation Learning Framework", "authors": [ "Tianshu Wu" ], "abstract": "Building humanoid robots capable of generalizable whole-body loco-manipulation in the real world remains a fundamental challenge.", "arxivId": "2605.20373v1", "url": "https://arxiv.org/abs/2605.20373v1", "pdfUrl": "https://arxiv.org/pdf/2605.20373", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.AI", "cs.CV" ], "publishedAt": "2026-05-19T18:24:05.000Z", "summary": "Building humanoid robots capable of generalizable whole-body loco-manipulation in the real world remains a fundamental challenge.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-teleoperation-robot-9f4f2b08:ln6L6IDgJZjOWCJodt", "title": "CEER: Compliant End-Effector and Root Control as a Unified Interface for Hierarchical Humanoid Loco-Manipulation", "authors": [ "Xinyuan Luo" ], "abstract": "Humanoid robots have achieved impressive locomotion performance, yet contact-rich and long-horizon manipulation remains a major bottleneck.", "arxivId": "2605.19981v1", "url": "https://arxiv.org/abs/2605.19981v1", "pdfUrl": "https://arxiv.org/pdf/2605.19981", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-19T15:23:40.000Z", "summary": "Humanoid robots have achieved impressive locomotion performance, yet contact-rich and long-horizon manipulation remains a major bottleneck.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Humanoid robots", "slug": "humanoid-robots", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-teleoperation-robot-9f4f2b08:jA2ABCUNmgYqzCyLo2", "title": "COBALT: Crowdsourcing Robot Learning via Cloud-Based Teleoperation with Smartphones", "authors": [ "Ayush Agarwal" ], "abstract": "The scarcity of large-scale, high-quality demonstration data remains a bottleneck in scaling imitation learning for robotic manipulation.", "arxivId": "2605.19138v2", "url": "https://arxiv.org/abs/2605.19138v2", "pdfUrl": "https://arxiv.org/pdf/2605.19138", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.AI", "cs.LG" ], "publishedAt": "2026-05-18T21:37:32.000Z", "summary": "The scarcity of large-scale, high-quality demonstration data remains a bottleneck in scaling imitation learning for robotic manipulation.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "AI infrastructure", "slug": "ai-infrastructure", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-teleoperation-robot-9f4f2b08:JA4wHN7eE0Dd9fFzpR", "title": "Hand-in-the-Loop: Improving VLA Policies for Dexterous Manipulation via Seamless Hand-Arm Intervention", "authors": [ "Zhuohang Li" ], "abstract": "Vision-Language-Action (VLA) models are prone to compounding errors in dexterous manipulation, where high-dimensional action spaces and contact-rich dynamics amplify small policy deviations over long horizons.", "arxivId": "2605.15157v2", "url": "https://arxiv.org/abs/2605.15157v2", "pdfUrl": "https://arxiv.org/pdf/2605.15157", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.LG" ], "publishedAt": "2026-05-14T17:51:40.000Z", "summary": "Vision-Language-Action (VLA) models are prone to compounding errors in dexterous manipulation, where high-dimensional action spaces and contact-rich dynamics amplify small policy deviations over long horizons.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-imitation-learning-robot-cca50e90:250JJ95tvLLqVkfdWh", "title": "DSSP: Diffusion State Space Policy with Full-History Encoding", "authors": [ "Zhiyuan Guan" ], "abstract": "Diffusion-based imitation learning has shown strong promise for robot manipulation.", "arxivId": "2605.14598v2", "url": "https://arxiv.org/abs/2605.14598v2", "pdfUrl": "https://arxiv.org/pdf/2605.14598", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-14T09:06:01.000Z", "summary": "Diffusion-based imitation learning has shown strong promise for robot manipulation.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-model-context-protocol-1cc7e65f:RxBFfVlGkwPpcTk8Nd", "title": "Content-Aware Attack Detection in LLM Agent Tool-Call Traffic: An Empirical Study of Features, Architectures, and Evaluation Protocols", "authors": [ "Sultan Zavrak" ], "abstract": "The Model Context Protocol (MCP) has become a widely adopted interface for LLM agents to invoke external tools, yet learned monitoring of MCP tool-call traffic remains underexplored.", "arxivId": "2605.11053v3", "url": "https://arxiv.org/abs/2605.11053v3", "pdfUrl": "https://arxiv.org/pdf/2605.11053", "categories": [ "robotics-research", "Robotics", "cs.CR", "cs.AI", "cs.LG" ], "publishedAt": "2026-05-11T14:55:48.000Z", "summary": "The Model Context Protocol (MCP) has become a widely adopted interface for LLM agents to invoke external tools, yet learned monitoring of MCP tool-call traffic remains underexplored.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-imitation-learning-robot-cca50e90:LSROiEWbOiHnEutbtT", "title": "How to Utilize Failure Demo Data?: Effective Data Selection for Imitation Learning Using Distribution Differences in Attention Mechanism", "authors": [ "Kana Miyamoto" ], "abstract": "Imitation learning for robotic tasks has relied primarily on policies trained only on successful demonstrations, although failures are unavoidable during human data collection.", "arxivId": "2605.07560v2", "url": "https://arxiv.org/abs/2605.07560v2", "pdfUrl": "https://arxiv.org/pdf/2605.07560", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-05-08T10:34:11.000Z", "summary": "Imitation learning for robotic tasks has relied primarily on policies trained only on successful demonstrations, although failures are unavoidable during human data collection.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-embodied-ai-00b1f9b2:hFj5gYWRJNYTvrphZA", "title": "VGGT-Segmentor: Geometry-Enhanced Cross-View Segmentation", "authors": [ "Yulu Gao" ], "abstract": "Instance-level object segmentation across disparate egocentric and exocentric views is a fundamental challenge in visual understanding, critical for applications in embodied AI and remote collaboration.", "arxivId": "2604.13596v3", "url": "https://arxiv.org/abs/2604.13596v3", "pdfUrl": "https://arxiv.org/pdf/2604.13596", "categories": [ "robotics-research", "Robotics", "cs.CV" ], "publishedAt": "2026-04-15T08:05:13.000Z", "summary": "Instance-level object segmentation across disparate egocentric and exocentric views is a fundamental challenge in visual understanding, critical for applications in embodied AI and remote collaboration.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-embodied-ai-00b1f9b2:M5vUlOMCRPTGF4Mgs8", "title": "Toward Self-Organizing Production Logistics: A Multi-Agent Approach", "authors": [ "Jan-Felix Klein" ], "abstract": "Production logistics (PL) is increasingly exposed to variability, dynamic interdependencies, and operational disturbances that challenge conventional centralized planning and control.", "arxivId": "2604.04753v2", "url": "https://arxiv.org/abs/2604.04753v2", "pdfUrl": "https://arxiv.org/pdf/2604.04753", "categories": [ "robotics-research", "Robotics", "eess.SY" ], "publishedAt": "2026-04-06T15:21:02.000Z", "summary": "Production logistics (PL) is increasingly exposed to variability, dynamic interdependencies, and operational disturbances that challenge conventional centralized planning and control.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-model-context-protocol-1cc7e65f:mY0WWY233Ei2_MEtGr", "title": "MatClaw: An Autonomous Code-First LLM Agent for End-to-End Materials Exploration", "authors": [ "Chenmu Zhang" ], "abstract": "Existing LLM agents for computational materials science are constrained by pipeline-bounded architectures tied to specific simulation codes and by dependence on manually written tool functions that grow with task scope.", "arxivId": "2604.02688v3", "url": "https://arxiv.org/abs/2604.02688v3", "pdfUrl": "https://arxiv.org/pdf/2604.02688", "categories": [ "robotics-research", "Robotics", "cond-mat.mtrl-sci", "cs.SE" ], "publishedAt": "2026-04-03T03:32:15.000Z", "summary": "Existing LLM agents for computational materials science are constrained by pipeline-bounded architectures tied to specific simulation codes and by dependence on manually written tool functions that grow with task scope.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-ro-468314c4:D0vANZFTt5USZn3EQq", "title": "Optimal Solutions for the Moving Target Vehicle Routing Problem with Obstacles via Lazy Branch and Price", "authors": [ "Anoop Bhat" ], "abstract": "The Moving Target Vehicle Routing Problem with Obstacles (MT-VRP-O) seeks trajectories for several agents that collectively intercept a set of moving targets.", "arxivId": "2603.21880v4", "url": "https://arxiv.org/abs/2603.21880v4", "pdfUrl": "https://arxiv.org/pdf/2603.21880", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2026-03-23T12:10:12.000Z", "summary": "The Moving Target Vehicle Routing Problem with Obstacles (MT-VRP-O) seeks trajectories for several agents that collectively intercept a set of moving targets.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-mobile-manipulation-ec740706:GYKLULciLPcLhnxII4", "title": "CritiSense: Critical Digital Literacy and Resilience Against Misinformation", "authors": [ "Firoj Alam" ], "abstract": "Misinformation on social media undermines informed decision-making and public trust.", "arxivId": "2603.16672v2", "url": "https://arxiv.org/abs/2603.16672v2", "pdfUrl": "https://arxiv.org/pdf/2603.16672", "categories": [ "robotics-research", "Robotics", "cs.AI", "cs.CL", "cs.CY" ], "publishedAt": "2026-03-17T15:37:49.000Z", "summary": "Misinformation on social media undermines informed decision-making and public trust.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-foundation-model-0634b051:r5ECtoheLDThDhiF8g", "title": "Investigating Robot Control Policy Learning for Autonomous X-ray-guided Spine Procedures", "authors": [ "Florence Klitzner" ], "abstract": "Imitation learning-based robot control policies are enjoying renewed interest in video-based robotics.", "arxivId": "2511.03882v2", "url": "https://arxiv.org/abs/2511.03882v2", "pdfUrl": "https://arxiv.org/pdf/2511.03882", "categories": [ "robotics-research", "Robotics", "cs.CV", "cs.AI", "cs.LG" ], "publishedAt": "2025-11-05T22:00:48.000Z", "summary": "Imitation learning-based robot control policies are enjoying renewed interest in video-based robotics.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-model-context-protocol-1cc7e65f:JassECz3OW1oiy4EBh", "title": "Ax-Prover: A Deep Reasoning Agentic Framework for Theorem Proving in Mathematics and Quantum Physics", "authors": [ "Benjamin Breen" ], "abstract": "We present Ax-Prover, a multi-agent system for automated theorem proving in Lean that can solve problems across diverse scientific domains and operate either autonomously or collaboratively with human experts.", "arxivId": "2510.12787v4", "url": "https://arxiv.org/abs/2510.12787v4", "pdfUrl": "https://arxiv.org/pdf/2510.12787", "categories": [ "robotics-research", "Robotics", "cs.AI", "cs.MA" ], "publishedAt": "2025-10-14T17:57:04.000Z", "summary": "We present Ax-Prover, a multi-agent system for automated theorem proving in Lean that can solve problems across diverse scientific domains and operate either autonomously or collaboratively with human experts.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-manipulation-4d4e16c1:qBIr7jJyuoAJcUAmGW", "title": "USIM and U0: A Vision-Language-Action Dataset and Model for General Underwater Robots", "authors": [ "Junwen Gu" ], "abstract": "Underwater environments pose unique challenges for robotic navigation and manipulation.", "arxivId": "2510.07869v4", "url": "https://arxiv.org/abs/2510.07869v4", "pdfUrl": "https://arxiv.org/pdf/2510.07869", "categories": [ "robotics-research", "Robotics", "cs.RO" ], "publishedAt": "2025-10-09T07:19:29.000Z", "summary": "Underwater environments pose unique challenges for robotic navigation and manipulation.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-robot-manipulation-4d4e16c1:L1fZYgSCjACzXxMQ5N", "title": "GAF: Gaussian Action Field as a 4D Representation for Dynamic World Modeling in Robotic Manipulation", "authors": [ "Ying Chai" ], "abstract": "Accurate scene perception is critical for vision-based robotic manipulation.", "arxivId": "2506.14135v5", "url": "https://arxiv.org/abs/2506.14135v5", "pdfUrl": "https://arxiv.org/pdf/2506.14135", "categories": [ "robotics-research", "Robotics", "cs.RO", "cs.CV" ], "publishedAt": "2025-06-17T02:55:20.000Z", "summary": "Accurate scene perception is critical for vision-based robotic manipulation.", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-model-context-protocol-1cc7e65f:9vdWz7A7HEh1AVIoT_", "title": "Survey of LLM Agent Communication with MCP: A Software Design Pattern Centric Review", "authors": [ "Anjana Sarkar" ], "abstract": "This survey investigates how classical software design patterns can enhance the reliability and scalability of communication in Large Language Model (LLM)-driven agentic AI systems, focusing particularly on the Model Context Protocol (MCP).", "arxivId": "2506.05364v2", "url": "https://arxiv.org/abs/2506.05364v2", "pdfUrl": "https://arxiv.org/pdf/2506.05364", "categories": [ "robotics-research", "Robotics", "cs.SE" ], "publishedAt": "2025-05-26T09:11:17.000Z", "summary": "This survey investigates how classical software design patterns can enhance the reliability and scalability of communication in Large Language Model (LLM)-driven agentic AI systems, focusing particularly on the Model Context Protocol (MCP).", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:dr5a7uljsA7Zn6hZZr", "title": "BOHM: Zero-Cost Hierarchical Attribution for Compound AI Systems", "authors": [ "Joss Armstrong" ], "abstract": "Compound AI systems route tasks through hierarchies of specialised components.", "arxivId": "2605.22866", "url": "https://arxiv.org/abs/2605.22866", "pdfUrl": "https://arxiv.org/pdf/2605.22866", "categories": [ "research", "Research", "cs.AI", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Compound AI systems route tasks through hierarchies of specialised components.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:cdbph9YoT39Q5eO_Sd", "title": "NeuroNL2LTL: A Neurosymbolic Framework for Natural Language Translation of Linear Temporal Logic", "authors": [ "Paapa Kwesi Quansah, Ernest Bonnah" ], "abstract": "Effectively translating between natural language (NL) and formal logics like Linear Temporal Logic (LTL) requires expertise that limits formal verification's reach in safety-critical development.", "arxivId": "2605.22874", "url": "https://arxiv.org/abs/2605.22874", "pdfUrl": "https://arxiv.org/pdf/2605.22874", "categories": [ "research", "Research", "cs.AI", "cs.LO" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Effectively translating between natural language (NL) and formal logics like Linear Temporal Logic (LTL) requires expertise that limits formal verification's reach in safety-critical development.", "relevanceScore": 0.59, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:9DeODrKDOZl32TLW8l", "title": "RMA: an Agentic System for Research-Level Mathematical Problems", "authors": [ "Zelin Zhao, Bo Yuan, Jaemoo Choi, Yongxin Chen" ], "abstract": "We present $\\textbf{Research Math Agents (RMA)}$, an agentic framework for automated reasoning on research-level mathematical problems.", "arxivId": "2605.22875", "url": "https://arxiv.org/abs/2605.22875", "pdfUrl": "https://arxiv.org/pdf/2605.22875", "categories": [ "research", "Research", "cs.AI", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "We present $\\textbf{Research Math Agents (RMA)}$, an agentic framework for automated reasoning on research-level mathematical problems.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:fvAoUvBXqolVmLVszs", "title": "SciAtlas: A Large-Scale Knowledge Graph for Automated Scientific Research", "authors": [ "Shuofei Qiao, Yunxiang Wei, Jiazheng Fan, Bin Wu, Busheng Zhang, Mengru Wang, Yuqi Zhu, Ningyu Zhang, Keyan Ding, Qiang Zhang, Huajun Chen" ], "abstract": "The exponential growth of global academic output has confronted researchers and AI agents with an unprecedented ``information explosion,'' where fragmented and unstructured knowledge organization impedes deep interdisciplinary integration.", "arxivId": "2605.22878", "url": "https://arxiv.org/abs/2605.22878", "pdfUrl": "https://arxiv.org/pdf/2605.22878", "categories": [ "research", "Research", "cs.AI", "cs.CL", "cs.IR" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "The exponential growth of global academic output has confronted researchers and AI agents with an unprecedented ``information explosion,'' where fragmented and unstructured knowledge organization impedes deep interdisciplinary integration.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:XoxUt_ekVryCkX1Dew", "title": "Energy per Successful Goal: Goal-Level Energy Accounting for Agentic AI Systems", "authors": [ "Deepak Panigrahy, Aakash Tyagi" ], "abstract": "Current AI energy benchmarks measure consumption at the granularity of a single model invocation or training run.", "arxivId": "2605.22883", "url": "https://arxiv.org/abs/2605.22883", "pdfUrl": "https://arxiv.org/pdf/2605.22883", "categories": [ "research", "Research", "cs.AI", "cs.LG", "cs.PF" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Current AI energy benchmarks measure consumption at the granularity of a single model invocation or training run.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:2CFUC0EEVcWW0368rw", "title": "ImProver 2: Iteratively Self-Improving LMs for Neurosymbolic Proof Optimization", "authors": [ "Riyaz Ahuja, Tate Rowney, Jeremy Avigad, Sean Welleck" ], "abstract": "Formal mathematics libraries are rapidly expanding, creating a growing need to refactor verified proofs for maintainability and to improve training data quality for neural provers.", "arxivId": "2605.22885", "url": "https://arxiv.org/abs/2605.22885", "pdfUrl": "https://arxiv.org/pdf/2605.22885", "categories": [ "research", "Research", "cs.AI", "cs.CL", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Formal mathematics libraries are rapidly expanding, creating a growing need to refactor verified proofs for maintainability and to improve training data quality for neural provers.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:YkSVeTMAm5JH3HBkYs", "title": "Mediative Fuzzy Logic: From Type-1 Foundations to Type-2, Type-3 and Quantum Extensions", "authors": [ "Oscar Montiel Ross" ], "abstract": "Mediative Fuzzy Logic was conceived as a practical scheme for reconciling hesitant or conflicting assessments in fuzzy control and decision-making.", "arxivId": "2605.22900", "url": "https://arxiv.org/abs/2605.22900", "pdfUrl": "https://arxiv.org/pdf/2605.22900", "categories": [ "research", "Research", "cs.AI", "cs.LO", "quant-ph" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Mediative Fuzzy Logic was conceived as a practical scheme for reconciling hesitant or conflicting assessments in fuzzy control and decision-making.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:DVW8curEITfwwzHK_o", "title": "EVE-Agent: Evidence-Verifiable Self-Evolving Agents", "authors": [ "Yamato Arai, Yuma Ichikawa" ], "abstract": "Self-evolving agents should not train on examples they cannot justify.", "arxivId": "2605.22905", "url": "https://arxiv.org/abs/2605.22905", "pdfUrl": "https://arxiv.org/pdf/2605.22905", "categories": [ "research", "Research", "cs.AI", "cs.CL" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Self-evolving agents should not train on examples they cannot justify.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:OBW5ODGuxK8245UJPH", "title": "The Deterministic Horizon: Impossibility Results as Design Specifications for Trustworthy AI Systems", "authors": [ "Dongxin Guo" ], "abstract": "Large language models now write software, draft legal documents, and produce clinical notes, yet fundamental limits, from Turing and Arrow to the No Free Lunch theorems, shape what computation can do.", "arxivId": "2605.23024", "url": "https://arxiv.org/abs/2605.23024", "pdfUrl": "https://arxiv.org/pdf/2605.23024", "categories": [ "research", "Research", "cs.AI", "cs.CC", "cs.CL" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Large language models now write software, draft legal documents, and produce clinical notes, yet fundamental limits, from Turing and Arrow to the No Free Lunch theorems, shape what computation can do.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ai:sOVRV_bmc9xGsTBwnF", "title": "PathCal: State-Aware Reflection-Marker Calibration for Efficient Reasoning", "authors": [ "Lingyu Jiang, Zirui Li, Shuo Xing, Peiran Li, Tsubasa Takahashi, Dengzhe Hou, Zhengzhong Tu, Kazunori Yamada, Fangzhou Lin" ], "abstract": "The emergence of Large Reasoning Language Models (LRMs) has paved the way for tackling complex reasoning tasks through test-time scaling by generating long-form Chain-of-Thought (CoT) trajectories during inference.", "arxivId": "2605.23074", "url": "https://arxiv.org/abs/2605.23074", "pdfUrl": "https://arxiv.org/pdf/2605.23074", "categories": [ "research", "Research", "cs.AI" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "The emergence of Large Reasoning Language Models (LRMs) has paved the way for tackling complex reasoning tasks through test-time scaling by generating long-form Chain-of-Thought (CoT) trajectories during inference.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv AI", "slug": "arxiv-ai", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "AI infrastructure", "slug": "ai-infrastructure", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:CWyHkmjhaBew7diAWF", "title": "Latent Cache Flow: Model-to-Model Communication Without Text", "authors": [ "Maximillian Rossi, Prajwal Raghunath, Eugene Wu" ], "abstract": "LLM agents today communicate via text, which incurs considerable latency and information loss due to the need to autoregressively decode the sharer model's state and encode at the receiver model.", "arxivId": "2605.22863", "url": "https://arxiv.org/abs/2605.22863", "pdfUrl": "https://arxiv.org/pdf/2605.22863", "categories": [ "research", "Research", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "LLM agents today communicate via text, which incurs considerable latency and information loss due to the need to autoregressively decode the sharer model's state and encode at the receiver model.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:FX1N_qlB8PzA1tDe67", "title": "Reading Calibrated Uncertainty from Language Model Trajectories", "authors": [ "Aliai Eusebi, Alexander Herzog, Xiaoyu Liang, Marie Vasek, Enrico Mariconti, Lorenzo Cavallaro" ], "abstract": "The maximum softmax probability (MSP) represents a default approach when evaluating uncertainty quantification for language model generation with structured output.", "arxivId": "2605.22864", "url": "https://arxiv.org/abs/2605.22864", "pdfUrl": "https://arxiv.org/pdf/2605.22864", "categories": [ "research", "Research", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "The maximum softmax probability (MSP) represents a default approach when evaluating uncertainty quantification for language model generation with structured output.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:7oUCXlglszjVZ8qwU1", "title": "FusionSense: Tri-Stage Near-Sensor Learning for Runtime-Adaptive Multimodal Edge Intelligence", "authors": [ "Sanggeon Yun, Ryozo Masukawa, Minhyoung Na, Hyunwoo Oh, Yoshiki Yamaguchi, Wenjun Huang, SungHeon Jeong, Mohsen Imani" ], "abstract": "Autonomous systems and smart-industry deployments increasingly split computation across near-sensor, edge, and cloud resources, where tight energy, latency, and reliability budgets demand run-time adaptivity.", "arxivId": "2605.22868", "url": "https://arxiv.org/abs/2605.22868", "pdfUrl": "https://arxiv.org/pdf/2605.22868", "categories": [ "research", "Research", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Autonomous systems and smart-industry deployments increasingly split computation across near-sensor, edge, and cloud resources, where tight energy, latency, and reliability budgets demand run-time adaptivity.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "AI infrastructure", "slug": "ai-infrastructure", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:9T9TJVSVeyxP0NaTuK", "title": "FuRA: Full-Rank Parameter-Efficient Fine-Tuning with Spectral Preconditioning", "authors": [ "Yequan Zhao, Ruijie Zhang, Liyan Tan, Niall Moran, Tong Qin, Zheng Zhang" ], "abstract": "Both full fine-tuning (Full FT) and parameter-efficient fine-tuning methods such as LoRA introduce weight updates without accounting for the spectral structure established during pretraining.", "arxivId": "2605.22869", "url": "https://arxiv.org/abs/2605.22869", "pdfUrl": "https://arxiv.org/pdf/2605.22869", "categories": [ "research", "Research", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Both full fine-tuning (Full FT) and parameter-efficient fine-tuning methods such as LoRA introduce weight updates without accounting for the spectral structure established during pretraining.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:Qsgu3FNsyTghKg9nFV", "title": "The Readout Shortcut: Positional Number Copying Dominates Arithmetic CoT Readout in Small Language Models", "authors": [ "Ming Liu" ], "abstract": "Chain-of-thought (CoT) prompting is necessary for arithmetic in small language models, yet shuffling its steps preserves most performance.", "arxivId": "2605.22870", "url": "https://arxiv.org/abs/2605.22870", "pdfUrl": "https://arxiv.org/pdf/2605.22870", "categories": [ "research", "Research", "cs.LG", "cs.AI", "cs.CL" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Chain-of-thought (CoT) prompting is necessary for arithmetic in small language models, yet shuffling its steps preserves most performance.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:k_2hg7Z-GDgfwfrIFk", "title": "Approximate Machine Unlearning through Manifold Representation Forgetting Guided by Self Mode Connectivity", "authors": [ "Weiqi Wang, Zhiyi Tian, Chenhan Zhang, Luoyu Chen, Shui Yu" ], "abstract": "Machine unlearning is a fundamental mechanism that enforces the right to be forgotten.", "arxivId": "2605.22871", "url": "https://arxiv.org/abs/2605.22871", "pdfUrl": "https://arxiv.org/pdf/2605.22871", "categories": [ "research", "Research", "cs.LG", "cs.AI", "stat.ML" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Machine unlearning is a fundamental mechanism that enforces the right to be forgotten.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:ai-WkDBClr0CJV0aj0", "title": "MedExpMem: Adapting Experience Memory for Differential Diagnosis", "authors": [ "Qianhan Feng, Zhongzhen Huang, Yakun Zhu, Yannian Gu, Winnie Chiu Wing Chu, Xiaofan Zhang, Qi Dou" ], "abstract": "Experienced physicians develop diagnostic expertise through clinical practice, acquiring not only disease knowledge but also the ability to differentiate confusable conditions.", "arxivId": "2605.22872", "url": "https://arxiv.org/abs/2605.22872", "pdfUrl": "https://arxiv.org/pdf/2605.22872", "categories": [ "research", "Research", "cs.LG", "cs.AI", "cs.CV" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Experienced physicians develop diagnostic expertise through clinical practice, acquiring not only disease knowledge but also the ability to differentiate confusable conditions.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:47mvqmE7wUqTkShF4L", "title": "When Do LLMs Reason? A Dynamical Systems View via Entropy Phase Transitions", "authors": [ "Wei Xia, Haoqing Wang, Zhi-Hong Deng, Yehui Tang" ], "abstract": "Chain-of-thought (CoT) reasoning has become the default strategy for enhancing LLM capabilities, yet its application raises a fundamental question: when is explicit reasoning actually beneficial?", "arxivId": "2605.22873", "url": "https://arxiv.org/abs/2605.22873", "pdfUrl": "https://arxiv.org/pdf/2605.22873", "categories": [ "research", "Research", "cs.LG", "cs.AI", "cs.CL" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Chain-of-thought (CoT) reasoning has become the default strategy for enhancing LLM capabilities, yet its application raises a fundamental question: when is explicit reasoning actually beneficial?", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:O-AKhhFMGkLCmDhMRb", "title": "WeCon: An Efficient Weight-Conditioned Neural Solver for Multi-Objective Combinatorial Optimization Problems", "authors": [ "Xuan Wu, Jinbiao Chen, Yang Li, Lijie Wen, Chunguo Wu, Yuanshu Li, Yubin Xiao, Chunyan Miao, You Zhou, Di Wang" ], "abstract": "Existing neural solvers for Multi-Objective Combinatorial Optimization Problems (MOCOPs) commonly adopt decomposition-based strategies that scalarize an MOCOP into multiple subproblems associated with distinct weight vectors.", "arxivId": "2605.22876", "url": "https://arxiv.org/abs/2605.22876", "pdfUrl": "https://arxiv.org/pdf/2605.22876", "categories": [ "research", "Research", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Existing neural solvers for Multi-Objective Combinatorial Optimization Problems (MOCOPs) commonly adopt decomposition-based strategies that scalarize an MOCOP into multiple subproblems associated with distinct weight vectors.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-ml:GSkX-9CJXw1jgGKLk-", "title": "Tensor Cache: Eviction-conditioned Associative Memory for Transformers", "authors": [ "Kabir Swain, Sijie Han, Daniel Karl I. Weidele, Mauro Martino, Antonio Torralba" ], "abstract": "Autoregressive Transformer KV caches grow linearly with context length; sliding-window caching bounds memory but discards evicted tokens entirely, so relevant evidence outside the window becomes inaccessible.", "arxivId": "2605.22884", "url": "https://arxiv.org/abs/2605.22884", "pdfUrl": "https://arxiv.org/pdf/2605.22884", "categories": [ "research", "Research", "cs.LG", "cs.AI" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Autoregressive Transformer KV caches grow linearly with context length; sliding-window caching bounds memory but discards evicted tokens entirely, so relevant evidence outside the window becomes inaccessible.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv ML", "slug": "arxiv-ml", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:4SLTTCvKqie6Z1jiZM", "title": "Remote Teleoperation of Endovascular Intervention Robots: A Systematic Review", "authors": [ "Xingyu Chen, Yinchao Yang, Nikola Fischer, Harry Robertshaw, Benjamin Jackson, Mohammad Shikh-Bahaei, Christos Bergeles, Thomas C Booth" ], "abstract": "Remote robotic-assisted endovascular intervention offers a promising approach to reduce clinician radiation exposure and physical strain, while extending specialized vascular care to geographically distant regions.", "arxivId": "2605.22889", "url": "https://arxiv.org/abs/2605.22889", "pdfUrl": "https://arxiv.org/pdf/2605.22889", "categories": [ "robotics", "Robotics", "cs.RO" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Remote robotic-assisted endovascular intervention offers a promising approach to reduce clinician radiation exposure and physical strain, while extending specialized vascular care to geographically distant regions.", "relevanceScore": 0.72, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:S4hKWiT8fGF7EnyJju", "title": "Extending Deep Event Visual Odometry with Sparse Point-Cloud Export", "authors": [ "Alireza Safdari, Sajad Ashraf" ], "abstract": "Event cameras are well suited for visual odometry under high-speed motion and challenging lighting conditions due to their low latency, high temporal resolution, and high dynamic range.", "arxivId": "2605.22890", "url": "https://arxiv.org/abs/2605.22890", "pdfUrl": "https://arxiv.org/pdf/2605.22890", "categories": [ "robotics", "Robotics", "cs.RO", "cs.CV" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Event cameras are well suited for visual odometry under high-speed motion and challenging lighting conditions due to their low latency, high temporal resolution, and high dynamic range.", "relevanceScore": 0.56, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "AI infrastructure", "slug": "ai-infrastructure", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:XU-I4DjuzLp9fQXKsY", "title": "Agentic-VLA: Efficient Online Adaptation for Vision-Language-Action Models", "authors": [ "Ruofan Jin, Zaixi Zhang" ], "abstract": "Vision-Language-Action (VLA) models have emerged as a promising paradigm for robotic manipulation by leveraging pre-trained vision-language representations.", "arxivId": "2605.22896", "url": "https://arxiv.org/abs/2605.22896", "pdfUrl": "https://arxiv.org/pdf/2605.22896", "categories": [ "robotics", "Robotics", "cs.RO", "cs.AI", "cs.LG" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Vision-Language-Action (VLA) models have emerged as a promising paradigm for robotic manipulation by leveraging pre-trained vision-language representations.", "relevanceScore": 0.72, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:phlKm1TjJ5r5_WhJnL", "title": "Robots That Know What to Ask: Recovering Misaligned Rewards through Targeted Explanations", "authors": [ "Helena Merker, Nick Walker, Andreea Bobu" ], "abstract": "Learning reward functions from demonstrations assumes that demonstrations provide adequate supervision over all features -- or task-relevant aspects of behavior.", "arxivId": "2605.22986", "url": "https://arxiv.org/abs/2605.22986", "pdfUrl": "https://arxiv.org/pdf/2605.22986", "categories": [ "robotics", "Robotics", "cs.RO", "cs.AI", "cs.HC" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Learning reward functions from demonstrations assumes that demonstrations provide adequate supervision over all features -- or task-relevant aspects of behavior.", "relevanceScore": 0.56, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:DnTCZ-90Ut254QBN2o", "title": "Verified Task-Space Motion Planning Under Joint-Space Constraints", "authors": [ "Hanjiang Hu, Changliu Liu, Yebin Wang" ], "abstract": "Reactive task-space planners such as Bug2 operate with fixed Cartesian step sizes and are unaware of the manipulator's joint-angle limits.", "arxivId": "2605.22991", "url": "https://arxiv.org/abs/2605.22991", "pdfUrl": "https://arxiv.org/pdf/2605.22991", "categories": [ "robotics", "Robotics", "cs.RO" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Reactive task-space planners such as Bug2 operate with fixed Cartesian step sizes and are unaware of the manipulator's joint-angle limits.", "relevanceScore": 0.56, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:Rh6xjNy0KmkOuRM40X", "title": "PIMbot: A Self-Adaptive Attack Framework for Adversarial Manipulation of Multi-Robot Reinforcement Learning", "authors": [ "Zexin Li, Ziliang Zhang, Hyoseung Kim, Cong Liu" ], "abstract": "Recent research has demonstrated the potential of reinforcement learning in effective multi-robot collaboration, particularly in social dilemmas where robots face a trade-off between self-interest and collective benefits.", "arxivId": "2605.23027", "url": "https://arxiv.org/abs/2605.23027", "pdfUrl": "https://arxiv.org/pdf/2605.23027", "categories": [ "robotics", "Robotics", "cs.RO" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Recent research has demonstrated the potential of reinforcement learning in effective multi-robot collaboration, particularly in social dilemmas where robots face a trade-off between self-interest and collective benefits.", "relevanceScore": 0.9, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:jwmgmu16tN0IkdnLP4", "title": "UfM*: Uncertainty from Motion* for DNN Depth Estimation Using Gaussians", "authors": [ "Soumya Sudhakar, Sertac Karaman, Vivienne Sze" ], "abstract": "Reliable uncertainty estimation is critical for deploying monocular depth deep neural networks (DNNs) in safety-critical robotic systems.", "arxivId": "2605.23098", "url": "https://arxiv.org/abs/2605.23098", "pdfUrl": "https://arxiv.org/pdf/2605.23098", "categories": [ "robotics", "Robotics", "cs.RO" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Reliable uncertainty estimation is critical for deploying monocular depth deep neural networks (DNNs) in safety-critical robotic systems.", "relevanceScore": 0.63, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "Robotics safety", "slug": "robotics-safety", "type": "topic" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:BfNHarxhfUrAO1-OM3", "title": "Four Simple Proprioceptive Estimators for Legged Robots", "authors": [ "Frank Dellaert, Chiyun Noh, Varun Agrawal, Ayoung Kim" ], "abstract": "Legged robots carry an IMU, but the inertial solution drifts because consumer-grade IMUs are noisy.", "arxivId": "2605.23100", "url": "https://arxiv.org/abs/2605.23100", "pdfUrl": "https://arxiv.org/pdf/2605.23100", "categories": [ "robotics", "Robotics", "cs.RO" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Legged robots carry an IMU, but the inertial solution drifts because consumer-grade IMUs are noisy.", "relevanceScore": 0.56, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:xRen8mPrJ7wa6Bboh7", "title": "$\\pi_0$-EqM: Equilibrium Matching for Closed-Loop Vision-Language-Action Control", "authors": [ "Huanming Liu, Congsheng Xu, Jianmin Ji, Yao Mu" ], "abstract": "Currently, Vision-Language-Action (VLA) models have become the most adopted paradigm for robotic manipulation for its great potential for task generalization.", "arxivId": "2605.23128", "url": "https://arxiv.org/abs/2605.23128", "pdfUrl": "https://arxiv.org/pdf/2605.23128", "categories": [ "robotics", "Robotics", "cs.RO" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "Currently, Vision-Language-Action (VLA) models have become the most adopted paradigm for robotic manipulation for its great potential for task generalization.", "relevanceScore": 0.72, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-robotics:AixxTFFbriSqVa6mSC", "title": "Semantic-Aware Guided Drone Exploration for Language-Conditioned 3D Indoor Mapping", "authors": [ "Nitin Vegesna, Avideh Zakhor" ], "abstract": "We present Semantic-Aware Guided Exploration, SAGE, a system for open-vocabulary exploration in unknown 3D indoor environments that preserves coverage-oriented behavior while allowing semantic cues to reprioritize frontier selection.", "arxivId": "2605.23160", "url": "https://arxiv.org/abs/2605.23160", "pdfUrl": "https://arxiv.org/pdf/2605.23160", "categories": [ "robotics", "Robotics", "cs.RO", "cs.CV" ], "publishedAt": "2026-05-25T04:00:00.000Z", "summary": "We present Semantic-Aware Guided Exploration, SAGE, a system for open-vocabulary exploration in unknown 3D indoor environments that preserves coverage-oriented behavior while allowing semantic cues to reprioritize frontier selection.", "relevanceScore": 0.72, "entities": [ { "name": "arXiv RO", "slug": "arxiv-ro", "type": "primary-source" }, { "name": "arXiv", "slug": "arxiv", "type": "research-source" }, { "name": "robotics research", "slug": "robotics-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-agent-benchmark-33922619:7VPggePK-xDttq_fGD", "title": "SkillOpt: Executive Strategy for Self-Evolving Agent Skills", "authors": [ "Yifan Yang" ], "abstract": "Agent skills today are hand-crafted, generated one-shot, or evolved through loosely controlled self-revision, none of which behaves like a deep-learning optimizer for the skill, and none of which reliably improves over its starting point under feedback.", "arxivId": "2605.23904v1", "url": "https://arxiv.org/abs/2605.23904v1", "pdfUrl": "https://arxiv.org/pdf/2605.23904", "categories": [ "agent-research", "Research", "cs.AI", "cs.CL" ], "publishedAt": "2026-05-22T17:59:50.000Z", "summary": "Agent skills today are hand-crafted, generated one-shot, or evolved through loosely controlled self-revision, none of which behaves like a deep-learning optimizer for the skill, and none of which reliably improves over its starting point under feedback.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-cv-5252b1e6:lE_g5_lwakmkW_nwOK", "title": "Geo-Align: Video Generation Alignment via Metric Geometry Reward", "authors": [ "Zizun Li" ], "abstract": "Camera-controlled video generation has achieved remarkable progress in recent years.", "arxivId": "2605.23903v1", "url": "https://arxiv.org/abs/2605.23903v1", "pdfUrl": "https://arxiv.org/pdf/2605.23903", "categories": [ "agent-research", "Research", "cs.CV" ], "publishedAt": "2026-05-22T17:59:43.000Z", "summary": "Camera-controlled video generation has achieved remarkable progress in recent years.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-cv-5252b1e6:zmfmhZp4iqrqLy6yRG", "title": "PiD: Fast and High-Resolution Latent Decoding with Pixel Diffusion", "authors": [ "Yifan Lu" ], "abstract": "Most practical high-resolution text-to-image systems, including latent diffusion and autoregressive models, perform generation in a compact latent space, and a decoder maps the generated latents back to pixels.", "arxivId": "2605.23902v1", "url": "https://arxiv.org/abs/2605.23902v1", "pdfUrl": "https://arxiv.org/pdf/2605.23902", "categories": [ "agent-research", "Research", "cs.CV" ], "publishedAt": "2026-05-22T17:59:42.000Z", "summary": "Most practical high-resolution text-to-image systems, including latent diffusion and autoregressive models, perform generation in a compact latent space, and a decoder maps the generated latents back to pixels.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-lg-6d63f1c1:5ru7qiloIL46ZUSDvN", "title": "LLMs as Noisy Channels: A Shannon Perspective on Model Capacity and Scaling Laws", "authors": [ "Xu Ouyang" ], "abstract": "Existing scaling laws for Large Language Models (LLMs), predominantly monotonic power laws, fail to explain emerging non-monotonic phenomena such as catastrophic overtraining and quantization-induced degradation, where performance deteriorates despite...", "arxivId": "2605.23901v1", "url": "https://arxiv.org/abs/2605.23901v1", "pdfUrl": "https://arxiv.org/pdf/2605.23901", "categories": [ "agent-research", "Research", "cs.LG", "cs.AI", "cs.IT" ], "publishedAt": "2026-05-22T17:59:38.000Z", "summary": "Existing scaling laws for Large Language Models (LLMs), predominantly monotonic power laws, fail to explain emerging non-monotonic phenomena such as catastrophic overtraining and quantization-induced degradation, where performance deteriorates despite...", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-ai-6dfa9da8:Ij_jipAsas0CbgeCtm", "title": "From Raw Experience to Skill Consumption: A Systematic Study of Model-Generated Agent Skills", "authors": [ "Zisu Huang" ], "abstract": "Language agents increasingly improve by reusing \\emph{skills} -- structured procedural artifacts distilled from past experience.", "arxivId": "2605.23899v1", "url": "https://arxiv.org/abs/2605.23899v1", "pdfUrl": "https://arxiv.org/pdf/2605.23899", "categories": [ "agent-research", "Research", "cs.AI" ], "publishedAt": "2026-05-22T17:59:12.000Z", "summary": "Language agents increasingly improve by reusing \\emph{skills} -- structured procedural artifacts distilled from past experience.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-vision-language-action-c9cae923:pqRIsGdq1tEu4FapR2", "title": "SPACENUM: Revisiting Spatial Numerical Understanding in VLMs", "authors": [ "Jianshu Zhang" ], "abstract": "Vision-Language Models (VLMs) are increasingly deployed in embodied environments, where they need produce numerical outputs such as action magnitudes and spatial coordinates.", "arxivId": "2605.23898v1", "url": "https://arxiv.org/abs/2605.23898v1", "pdfUrl": "https://arxiv.org/pdf/2605.23898", "categories": [ "agent-research", "Research", "cs.AI" ], "publishedAt": "2026-05-22T17:58:36.000Z", "summary": "Vision-Language Models (VLMs) are increasingly deployed in embodied environments, where they need produce numerical outputs such as action magnitudes and spatial coordinates.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-cv-5252b1e6:BWKU2plfMp7MhNzSH6", "title": "ETCHR: Editing To Clarify and Harness Reasoning", "authors": [ "Beichen Zhang" ], "abstract": "Multimodal Large Language Models have advanced visual reasoning, yet a purely textual chain of thought remains a bottleneck for questions that require fine-grained focus or view transformations.", "arxivId": "2605.23897v1", "url": "https://arxiv.org/abs/2605.23897v1", "pdfUrl": "https://arxiv.org/pdf/2605.23897", "categories": [ "agent-research", "Research", "cs.CV", "cs.AI", "cs.CL" ], "publishedAt": "2026-05-22T17:58:28.000Z", "summary": "Multimodal Large Language Models have advanced visual reasoning, yet a purely textual chain of thought remains a bottleneck for questions that require fine-grained focus or view transformations.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-cv-5252b1e6:1dzPmD5xjQJ1CPj1hV", "title": "From Activation to Causality: Discovery of Causal Visual Representations in the Human Brain", "authors": [ "Yuval Golbari" ], "abstract": "Identifying which brain regions represent a visual concept in the human brain is a central challenge in neuroscience.", "arxivId": "2605.23895v1", "url": "https://arxiv.org/abs/2605.23895v1", "pdfUrl": "https://arxiv.org/pdf/2605.23895", "categories": [ "agent-research", "Research", "cs.CV" ], "publishedAt": "2026-05-22T17:56:37.000Z", "summary": "Identifying which brain regions represent a visual concept in the human brain is a central challenge in neuroscience.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-lg-6d63f1c1:d2RTIweql7cmB-b09e", "title": "Complete-muE: Optimal Hyperparameter Transfer and Scaling for MoE Models", "authors": [ "Hongwu Peng" ], "abstract": "We propose Complete-muE, a framework which targets hyperparameter transfer across dense FFN and any Mixture-of-Experts (MoE) setups in transformer blocks.", "arxivId": "2605.23893v1", "url": "https://arxiv.org/abs/2605.23893v1", "pdfUrl": "https://arxiv.org/pdf/2605.23893", "categories": [ "agent-research", "Research", "cs.LG" ], "publishedAt": "2026-05-22T17:56:13.000Z", "summary": "We propose Complete-muE, a framework which targets hyperparameter transfer across dense FFN and any Mixture-of-Experts (MoE) setups in transformer blocks.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-lg-6d63f1c1:uLQic3kOqMNP_kkDIl", "title": "CHRONOS: Temporally-Aware Multi-Agent Coordination for Evolving Data Marketplaces", "authors": [ "Joydeep Chandra" ], "abstract": "Temporal knowledge-graph data marketplaces face three coupled failures in static designs: stale hybrid index shortcuts reduce recall as edges evolve, stationary Shapley pricing misattributes value after distribution shifts, and uncoordinated agents...", "arxivId": "2605.23887v1", "url": "https://arxiv.org/abs/2605.23887v1", "pdfUrl": "https://arxiv.org/pdf/2605.23887", "categories": [ "agent-research", "Research", "cs.DB", "cs.AI", "cs.CR" ], "publishedAt": "2026-05-22T17:47:45.000Z", "summary": "Temporal knowledge-graph data marketplaces face three coupled failures in static designs: stale hybrid index shortcuts reduce recall as edges evolve, stationary Shapley pricing misattributes value after distribution shifts, and uncoordinated agents...", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-lg-6d63f1c1:Y0z1O7qmWc9_3SHkkP", "title": "On the Stability of Spherical Hellinger-Kantorovich Flows and Their Implications for Differential Privacy", "authors": [ "Aratrika Mustafi" ], "abstract": "Gradient-flow sampling interprets a Gibbs distribution as the minimizer of an energy functional over probability measures and generates dynamics converging to this target.", "arxivId": "2605.23879v1", "url": "https://arxiv.org/abs/2605.23879v1", "pdfUrl": "https://arxiv.org/pdf/2605.23879", "categories": [ "agent-research", "Research", "stat.ML", "cs.CR", "cs.LG" ], "publishedAt": "2026-05-22T17:38:20.000Z", "summary": "Gradient-flow sampling interprets a Gibbs distribution as the minimizer of an energy functional over probability measures and generates dynamics converging to this target.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-agent-benchmark-33922619:2aW-c-3XuociQn7Dc7", "title": "Agentic Proving for Program Verification", "authors": [ "Alessandro Sosso" ], "abstract": "Agentic systems have recently emerged as state-of-the-art approaches for automated theorem proving in formal mathematics.", "arxivId": "2605.23772v1", "url": "https://arxiv.org/abs/2605.23772v1", "pdfUrl": "https://arxiv.org/pdf/2605.23772", "categories": [ "agent-research", "Research", "cs.AI", "cs.LO", "cs.PL" ], "publishedAt": "2026-05-22T15:41:27.000Z", "summary": "Agentic systems have recently emerged as state-of-the-art approaches for automated theorem proving in formal mathematics.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-agent-benchmark-33922619:EI9Ngal-x68u4BYdIj", "title": "PhotoFlow: Agentic 3D Virtual Photography Missions", "authors": [ "Jiarui Guo" ], "abstract": "Virtual photography asks an agent to enter a prepared 3D scene with no preselected camera pose or reference image, infer a suitable shot from scene information and a language intent, choose executable camera parameters, and render the final photograph.", "arxivId": "2605.23771v1", "url": "https://arxiv.org/abs/2605.23771v1", "pdfUrl": "https://arxiv.org/pdf/2605.23771", "categories": [ "agent-research", "Research", "cs.CV", "cs.AI", "cs.MA" ], "publishedAt": "2026-05-22T15:40:52.000Z", "summary": "Virtual photography asks an agent to enter a prepared 3D scene with no preselected camera pose or reference image, infer a suitable shot from scene information and a language intent, choose executable camera parameters, and render the final photograph.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-multi-agent-system-6f52d53f:WB--ZmPAFzlRfQoW8Z", "title": "SeedER: Seed-and-Expand Retrieval from Knowledge Graphs", "authors": [ "Hamed Shirzad" ], "abstract": "Knowledge graphs (KGs) offer a rich representation for relational knowledge, but their irregular structure makes retrieval challenging: ego-graph expansion grows rapidly, and dense embedding methods struggle with multi-hop compositional queries.", "arxivId": "2605.23753v1", "url": "https://arxiv.org/abs/2605.23753v1", "pdfUrl": "https://arxiv.org/pdf/2605.23753", "categories": [ "agent-research", "Research", "cs.LG" ], "publishedAt": "2026-05-22T15:26:31.000Z", "summary": "Knowledge graphs (KGs) offer a rich representation for relational knowledge, but their irregular structure makes retrieval challenging: ego-graph expansion grows rapidly, and dense embedding methods struggle with multi-hop compositional queries.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-agent-security-7627deac:2IGXtUkXLKeighY2TR", "title": "MemAudit: Post-hoc Auditing of Poisoned Agent Memory via Causal Attribution and Structural Anomaly Detection", "authors": [ "Zhewen Tan" ], "abstract": "Large language model agents increasingly rely on persistent memory to store past interactions, retrieve relevant demonstrations, and improve long-horizon task execution.", "arxivId": "2605.23723v1", "url": "https://arxiv.org/abs/2605.23723v1", "pdfUrl": "https://arxiv.org/pdf/2605.23723", "categories": [ "agent-research", "Research", "cs.AI" ], "publishedAt": "2026-05-22T15:03:13.000Z", "summary": "Large language model agents increasingly rely on persistent memory to store past interactions, retrieve relevant demonstrations, and improve long-horizon task execution.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-tool-use-agent-838879e2:GVoX8diVsbpxL741wv", "title": "RF Instrument Agent (RFIA): Empowering RF Instruments with Natural Language Understanding, Scheduling and Execution of Complex Tasks", "authors": [ "Chunhui Li" ], "abstract": "Modern radio-frequency (RF) instruments, such as vector network analyzers (VNAs), already provide mature remote-control interfaces.", "arxivId": "2605.23636v1", "url": "https://arxiv.org/abs/2605.23636v1", "pdfUrl": "https://arxiv.org/pdf/2605.23636", "categories": [ "agent-research", "Research", "eess.SY" ], "publishedAt": "2026-05-22T13:51:18.000Z", "summary": "Modern radio-frequency (RF) instruments, such as vector network analyzers (VNAs), already provide mature remote-control interfaces.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-vision-language-action-c9cae923:QscVYwgG1NOHa2N-fz", "title": "TactileReflex: Noise-Statistics-Driven Vision-Tactile Reflex Control for Force-Sensitive Manipulation", "authors": [ "Ziyan Feng" ], "abstract": "Manipulating fragile deformable containers, such as disposable plastic cups filled with liquid, demands real-time grip-force adaptation within an extremely narrow force margin: insufficient force causes slip, while excessive force irreversibly deforms the...", "arxivId": "2605.23568v1", "url": "https://arxiv.org/abs/2605.23568v1", "pdfUrl": "https://arxiv.org/pdf/2605.23568", "categories": [ "agent-research", "Research", "cs.RO", "eess.SY" ], "publishedAt": "2026-05-22T12:35:28.000Z", "summary": "Manipulating fragile deformable containers, such as disposable plastic cups filled with liquid, demands real-time grip-force adaptation within an extremely narrow force margin: insufficient force causes slip, while excessive force irreversibly deforms the...", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-multi-agent-system-6f52d53f:1PNxstfbNlA1RBq2_d", "title": "ARMS: Automatic Reward Shaping for Sparse-Reward Multi-Agent Reinforcement Learning", "authors": [ "Elie Abboud" ], "abstract": "Sparse rewards are a major bottleneck in multi-agent reinforcement learning (MARL), where simultaneous learning induces non-stationarity and makes reward design especially delicate.", "arxivId": "2605.23562v1", "url": "https://arxiv.org/abs/2605.23562v1", "pdfUrl": "https://arxiv.org/pdf/2605.23562", "categories": [ "agent-research", "Research", "cs.MA", "cs.AI" ], "publishedAt": "2026-05-22T12:29:29.000Z", "summary": "Sparse rewards are a major bottleneck in multi-agent reinforcement learning (MARL), where simultaneous learning induces non-stationarity and makes reward design especially delicate.", "relevanceScore": 0.68, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-tool-use-agent-838879e2:dSNHq5kqBQhitXbE5i", "title": "Goal-Conditioned Agents that Learn Everything All at Once", "authors": [ "Michael Matthews" ], "abstract": "A goal-conditioned reinforcement learning agent exploring an environment will see a wealth of information throughout a trajectory, most of which is discarded when only performing on-policy updates with respect to the commanded goal.", "arxivId": "2605.23551v1", "url": "https://arxiv.org/abs/2605.23551v1", "pdfUrl": "https://arxiv.org/pdf/2605.23551", "categories": [ "agent-research", "Research", "cs.LG", "cs.AI" ], "publishedAt": "2026-05-22T12:17:09.000Z", "summary": "A goal-conditioned reinforcement learning agent exploring an environment will see a wealth of information throughout a trajectory, most of which is discarded when only performing on-policy updates with respect to the commanded goal.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-agent-security-7627deac:lNif_cC7RRrEH14L74", "title": "LiveFigure: Generating Editable Scientific Illustration with VLM Agents", "authors": [ "Chenyang Shao" ], "abstract": "Scientific illustrations are essential for depicting conceptual designs, methodologies, and experimental workflows in research, playing a pivotal role in communicating complex academic insights.", "arxivId": "2605.23527v1", "url": "https://arxiv.org/abs/2605.23527v1", "pdfUrl": "https://arxiv.org/pdf/2605.23527", "categories": [ "agent-research", "Research", "cs.CE" ], "publishedAt": "2026-05-22T11:42:55.000Z", "summary": "Scientific illustrations are essential for depicting conceptual designs, methodologies, and experimental workflows in research, playing a pivotal role in communicating complex academic insights.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-vision-language-action-c9cae923:THafy2zr6TPOUbzD_x", "title": "DrawVideo: Generating Long Video from Storyboard Keyframe Sketches", "authors": [ "Chuanzhi Xu" ], "abstract": "Long video generation requires high-fidelity synthesis, coherent narrative structure, and user control over extended time spans.", "arxivId": "2605.23508v1", "url": "https://arxiv.org/abs/2605.23508v1", "pdfUrl": "https://arxiv.org/pdf/2605.23508", "categories": [ "agent-research", "Research", "cs.GR", "cs.AI", "cs.CV" ], "publishedAt": "2026-05-22T11:16:05.000Z", "summary": "Long video generation requires high-fidelity synthesis, coherent narrative structure, and user control over extended time spans.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-vision-language-action-c9cae923:NXa4x_BBqFfJShFNTZ", "title": "Semantically Structured Mixture-of-Experts for Compositional Robotic Manipulation", "authors": [ "Chengyu Deng" ], "abstract": "Diffusion-based policies have established a new standard for precise robotic manipulation but face a critical scalability bottleneck: high-performance models are computationally expensive, while lightweight alternatives often fail to generalize across diverse...", "arxivId": "2605.23477v1", "url": "https://arxiv.org/abs/2605.23477v1", "pdfUrl": "https://arxiv.org/pdf/2605.23477", "categories": [ "agent-research", "Research", "cs.RO" ], "publishedAt": "2026-05-22T10:38:59.000Z", "summary": "Diffusion-based policies have established a new standard for precise robotic manipulation but face a critical scalability bottleneck: high-performance models are computationally expensive, while lightweight alternatives often fail to generalize across...", "relevanceScore": 0.86, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-ai-agent-f8e6e527:D5ThZEXOoErn7Do6tw", "title": "AI Assurance: A Comprehensive Testing Strategy for Enterprise AI Systems", "authors": [ "Chitra Badagi" ], "abstract": "Enterprise AI systems, built on large language models, retrieval pipelines and autonomous agents, introduce a class of risks that traditional software quality assurance was never designed to address.", "arxivId": "2605.23459v1", "url": "https://arxiv.org/abs/2605.23459v1", "pdfUrl": "https://arxiv.org/pdf/2605.23459", "categories": [ "agent-research", "Research", "cs.SE", "cs.AI" ], "publishedAt": "2026-05-22T10:19:19.000Z", "summary": "Enterprise AI systems, built on large language models, retrieval pipelines and autonomous agents, introduce a class of risks that traditional software quality assurance was never designed to address.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-ai-agent-f8e6e527:3oj9p9mo3K_4LZL6sg", "title": "Socially fluent AI decouples conversational signals from source identity in online interaction", "authors": [ "Lixiang Yan" ], "abstract": "Socially fluent agentic AI can now participate in online interaction in ways that resemble ordinary human conversation, potentially weakening people's ability to infer who is human from conversational signals alone.", "arxivId": "2605.23426v1", "url": "https://arxiv.org/abs/2605.23426v1", "pdfUrl": "https://arxiv.org/pdf/2605.23426", "categories": [ "agent-research", "Research", "cs.HC", "cs.AI" ], "publishedAt": "2026-05-22T09:37:36.000Z", "summary": "Socially fluent agentic AI can now participate in online interaction in ways that resemble ordinary human conversation, potentially weakening people's ability to infer who is human from conversational signals alone.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-agent-security-7627deac:ke48PP9D6OYKRoEyEg", "title": "Security, Privacy, and Ethical Risks in OpenClaw", "authors": [ "Yutong Jin" ], "abstract": "This paper systematically investigates the security, privacy, and ethical risks, as well as the traceability challenges of OpenClaw, a locally executable AI agent system for natural language interaction and real-world task completion.", "arxivId": "2605.23330v1", "url": "https://arxiv.org/abs/2605.23330v1", "pdfUrl": "https://arxiv.org/pdf/2605.23330", "categories": [ "agent-research", "Research", "cs.CR" ], "publishedAt": "2026-05-22T07:45:04.000Z", "summary": "This paper systematically investigates the security, privacy, and ethical risks, as well as the traceability challenges of OpenClaw, a locally executable AI agent system for natural language interaction and real-world task completion.", "relevanceScore": 0.75, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-cat-cs-ai-6dfa9da8:PwLNayI8OVFT7TSksv", "title": "Is Capability a Liability? More Capable Language Models Make Worse Forecasts When It Matters Most", "authors": [ "Nick Merrill" ], "abstract": "We document inverse scaling in LLMs on forecasting problems whose underlying time series exhibit superlinear growth and tail risk of regime change, a structure common in finance and epidemiology.", "arxivId": "2605.22672v2", "url": "https://arxiv.org/abs/2605.22672v2", "pdfUrl": "https://arxiv.org/pdf/2605.22672", "categories": [ "agent-research", "Research", "cs.AI" ], "publishedAt": "2026-05-21T16:14:33.000Z", "summary": "We document inverse scaling in LLMs on forecasting problems whose underlying time series exhibit superlinear growth and tail risk of regime change, a structure common in finance and epidemiology.", "relevanceScore": 0.52, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" }, { "id": "arxiv-query-all-multi-agent-system-6f52d53f:Kz9_XzUkfqVTagZmPn", "title": "Boiling the Frog: A Multi-Turn Benchmark for Agentic Safety", "authors": [ "Piercosma Bisconti" ], "abstract": "Background. Traditional safety benchmarks for language models evaluate generated text: whether a model outputs toxic language, reproduces bias, or follows harmful instructions.", "arxivId": "2605.22643v2", "url": "https://arxiv.org/abs/2605.22643v2", "pdfUrl": "https://arxiv.org/pdf/2605.22643", "categories": [ "agent-research", "Research", "cs.CL" ], "publishedAt": "2026-05-21T15:50:18.000Z", "summary": "Background. Traditional safety benchmarks for language models evaluate generated text: whether a model outputs toxic language, reproduces bias, or follows harmful instructions.", "relevanceScore": 0.75, "entities": [ { "name": "arXiv", "slug": "arxiv", "type": "primary-source" }, { "name": "Autonomous agents", "slug": "autonomous-agents", "type": "topic" }, { "name": "model research", "slug": "model-research", "type": "surface" } ], "sourceStatus": "source-metadata-only" } ] }