机器人 / 具身智能

Leveraging Large (Visual) Language Models for Robot 3D Scene Understanding

William Chen, Siyi Hu, Rajat Talak, Luca Carlone

专题命中 具身推理 ：robotics(abstract)；分类 cs.RO、cs.CV、cs.LG

Comments arXiv admin note: text overlap with arXiv:2206.04585

Open-Vocabulary Affordance Detection in 3D Point Clouds

Toan Nguyen, Minh Nhat Vu, An Vuong, Dzung Nguyen, Thieu Vo, Ngan Le, Anh Nguyen

专题命中 具身推理 ：robotic(abstract)；分类 cs.RO、cs.AI、cs.CV

Comments Accepted at The 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2023)

Compositional 3D Human-Object Neural Animation

Zhi Hou, Baosheng Yu, Dacheng Tao

专题命中 具身推理 ：robotics(abstract)；分类 cs.AI、cs.CV、cs.LG

Comments 14 pages, 6 figures

Neural Groundplans: Persistent Neural Scene Representations from a Single Image

Prafull Sharma, Ayush Tewari, Yilun Du, Sergey Zakharov, Rares Ambrus, Adrien Gaidon, William T. Freeman, Fredo Durand, Joshua B. Tenenbaum, Vincent Sitzmann

专题命中 具身推理 ：robotics(abstract)；分类 cs.AI、cs.CV、cs.LG

Comments Project page: https://prafullsharma.net/neural_groundplans/

Meta-World Conditional Neural Processes

Suzan Ece Ada, Emre Ugur

专题命中 具身推理 ：world model(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments 11 pages, 9 figures

Learning to Forecast Aleatoric and Epistemic Uncertainties over Long Horizon Trajectories

Aastha Acharya, Rebecca Russell, Nisar R. Ahmed

专题命中 具身推理 ：world model(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments Accepted to ICRA 2023

Task-Aware Asynchronous Multi-Task Model with Class Incremental Contrastive Learning for Surgical Scene Understanding

Lalithkumar Seenivasan, Mobarakol Islam, Mengya Xu, Chwee Ming Lim, Hongliang Ren

专题命中 具身推理 ：robotic(abstract)；分类 cs.RO、cs.AI、cs.CV

Comments Manuscript accepted in the International Journal of Computer Assisted Radiology and Surgery. codes available: https://github.com/lalithjets/Domain-adaptation-in-MTL

Feature-Realistic Neural Fusion for Real-Time, Open Set Scene Understanding

Kirill Mazur, Edgar Sucar, Andrew J. Davison

专题命中 具身推理 ：robotics(abstract)；分类 cs.RO、cs.CV、cs.LG

Comments For our project page, see https://makezur.github.io/FeatureRealisticFusion/

Competency Assessment for Autonomous Agents using Deep Generative Models

Aastha Acharya, Rebecca Russell, Nisar R. Ahmed

专题命中 具身推理 ：world model(abstract)；分类 cs.RO、cs.AI、cs.LG

DiffSDFSim: Differentiable Rigid-Body Dynamics With Implicit Shapes

Michael Strecke, Joerg Stueckler

专题命中 具身推理 ：robotics(abstract)；分类 cs.RO、cs.CV、cs.LG

Comments 22 pages, 23 Figures (including supplementary material). Presented 3DV 2021. Project website: https://diffsdfsim.is.tue.mpg.de/

Journal ref 2021 International Conference on 3D Vision (3DV)

Commonsense Spatial Reasoning for Visually Intelligent Agents

Agnese Chiatti, Gianluca Bardaro, Enrico Motta, Enrico Daga

专题命中 具身推理 ：robotic(abstract)；分类 cs.RO、cs.AI、cs.CV

A Kinematic Bottleneck Approach For Pose Regression of Flexible Surgical Instruments directly from Images

Luca Sestini, Benoit Rosa, Elena De Momi, Giancarlo Ferrigno, Nicolas Padoy

专题命中 具身推理 ：robotic(abstract)；分类 cs.RO、cs.AI、cs.CV

Bridging Scene Understanding and Task Execution with Flexible Simulation Environments

Zachary Ravichandran, J. Daniel Griffith, Benjamin Smith, Costas Frost

专题命中 具身推理 ：robotics(abstract)；分类 cs.RO、cs.AI、cs.CV

Dream to Control: Learning Behaviors by Latent Imagination

Danijar Hafner, Timothy Lillicrap, Jimmy Ba, Mohammad Norouzi

专题命中 具身推理 ：world model(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments 9 pages, 12 figures

Learning by Abstraction: The Neural State Machine

Drew A. Hudson, Christopher D. Manning

专题命中 具身推理 ：world model(abstract)；分类 cs.AI、cs.CV、cs.LG

Comments Published as a conference paper at NeurIPS 2019 (spotlight)

Engineered Self-Organization for Resilient Robot Self-Assembly with Minimal Surprise

Tanja Katharina Kaiser, Heiko Hamann

专题命中 具身推理 ：world model(abstract)；robotic(abstract)

Self Training Autonomous Driving Agent

Shashank Kotyan, Danilo Vasconcellos Vargas, Venkanna U

专题命中 具身推理 ：world model(abstract)；分类 cs.RO、cs.AI、cs.CV

SWNet: Small-World Neural Networks and Rapid Convergence

Mojan Javaheripi, Bita Darvish Rouhani, Farinaz Koushanfar

专题命中 具身推理 ：world model(abstract)；分类 cs.AI、cs.CV、cs.LG

Scene Understanding Networks for Autonomous Driving based on Around View Monitoring System

JeongYeol Baek, Ioana Veronica Chelu, Livia Iordache, Vlad Paunescu, HyunJoo Ryu, Alexandru Ghiuta, Andrei Petreanu, YunSung Soh, Andrei Leica, ByeongMoon Jeon

专题命中 具身推理 ：navigation(abstract)；分类 cs.AI、cs.CV、cs.LG

Comments Accepted by CVPR 2018 Workshop on Autonomous Driving

Emergence of Structured Behaviors from Curiosity-Based Intrinsic Motivation

Nick Haber, Damian Mrowca, Li Fei-Fei, Daniel L. K. Yamins

专题命中 具身推理 ：world model(abstract)；分类 cs.AI、cs.CV、cs.LG

Comments 6 pages, 5 figures

Discovering Latent States for Model Learning: Applying Sensorimotor Contingencies Theory and Predictive Processing to Model Context

Nikolas J. Hemion

专题命中 具身推理 ：world model(abstract)；分类 cs.RO、cs.AI、cs.LG

Towards Holistic Scene Understanding: Feedback Enabled Cascaded Classification Models

Congcong Li, Adarsh Kowdle, Ashutosh Saxena, Tsuhan Chen

专题命中 具身推理 ：robotic(abstract)；分类 cs.RO、cs.AI、cs.CV

Comments 14 pages, 11 figures

Did the Grid Erase the Event? EndoClock for Auditing Medical World-Model Pipelines

网格是否抹去了事件？用于审计医疗世界模型流程的EndoClock

Yarin Udi, Tom Sharon-Shahak, Roee Masad, Dan Pri-Tal

机构 * Diastole Medical R&D（Diastole医疗研发机构） 纠错

专题命中 具身推理 ：world model(abstract,comments)；分类 cs.CV、cs.LG

AI总结 本研究针对医疗世界模型同步预处理可能抹去任务相关证据的问题，提出EndoClock审计方法及四分类法，以超声心动图为例验证其有效性，为医疗AI流程审计提供可执行方案。

Comments Accepted for publication at the 1st MICCAI Workshop on Medical World Models (MWM 2026)

Technical Report for ICRA 2026 GOOSE 2D Fine-Grained Semantic Segmentation Challenge: Exploring Query-Based Segmentation and Increased Spatial Context for Outdoor Scene Understanding

ICRA 2026 GOOSE 2D细粒度语义分割挑战赛技术报告：探索基于查询的分割和增加空间上下文用于户外场景理解

David Pascual-Hernández, Roberto Calvo-Palomino, Inmaculada Mora-Jiménez, Jose María Cañas-Plaza

机构 * Rey Juan Carlos University（胡安卡洛斯国王大学） 纠错

专题命中 具身推理 ：robotics(abstract,comments)；分类 cs.RO、cs.CV

AI总结 本报告提出基于SegFormer和Mask2Former的细粒度语义分割方法，通过增大训练裁剪尺寸和测试时增强，在GOOSE挑战赛上达到69.6% mIoU，验证了查询式分割和空间上下文的重要性。

Comments Ranked 5th in the GOOSE 2D Fine-Grained Semantic Segmentation Challenge at the IEEE ICRA 2026 Workshop on Field Robotics

Conditional Affordance Learning for Driving in Urban Environments

面向城市环境的条件性 affordance 学习

Axel Sauer, Nikolay Savinov, Andreas Geiger

机构 * Chair of Robotics Science and System Intelligence, Technical University of Munich（慕尼黑技术大学机器人科学与系统智能系） 纠错

专题命中 具身推理 ：navigation(abstract)；分类 cs.RO、cs.LG；robot learning(comments)

AI总结 本文提出了一种直接感知方法，通过将视频输入映射到适合复杂城市环境自主导航的中间表示，结合高层方向输入，实现了比现有强化学习和条件模仿学习方法更高的目标导向导航性能，并首次通过图像级标签处理交通灯和速度标志，显著减少模拟中的交通事故。

Comments Accepted for Conference on Robot Learning (CoRL) 2018

Affordance-Based Disambiguation of Surgical Instructions for Collaborative Robot-Assisted Surgery

Ana Davila, Jacinto Colan, Yasuhisa Hasegawa

机构 * Nagoya University, Japan（名古屋大学） 纠错

专题命中 具身推理 ：robotic(abstract,comments)；分类 cs.RO；robot learning(comments)

Comments To be presented at the 1st Workshop on Intelligent Cobodied Assistance and Robotic Empowerment (iCARE). 2025 Conference on Robot Learning (CoRL)

Marionette: Predicting World States, Rendering Geometry, Painting Appearance

Marionette：预测世界状态、渲染几何、绘制外观

Zian Meng, Zhen Li, Chuanhao Li, Qiang Li, Kaipeng Zhang

机构 * Alaya Lab（Alaya实验室） ； Shanghai Innovation Institute（上海创新研究院） ； Huazhong University of Science and Technology（华中科技大学） 纠错

专题命中 具身推理 ：world model(abstract)；分类 cs.AI、cs.CV

AI总结 Marionette是面向带关节角色的交互式游戏的世界模型，通过显式建模3D世界状态、委托几何计算、合成外观，实现了可控的长时序行为，且外观生成无明显保真度损失。

Comments Project page: https://alayalab.github.io/Marionette/

Semantic Radiance Fields as Simulators for Spatial Reasoning in Real-World Scenes

语义辐射场作为真实场景空间推理的模拟器

Nico Heider, Michał Jan Włodarczyk, Katarzyna Wasielewska-Michniewska, Przemysław Hołda, Martin Schieck, Marcin Paprzycki, Maria Ganzha, Bogdan Franczyk

机构 * Leipzig University（莱比锡大学） ； Systems Research Institute Polish Academy of Sciences（波兰科学院系统研究所） ； Wrocław University of Economics（弗罗茨瓦夫经济大学） 纠错

专题命中 具身推理 ：embodied agent(abstract)；分类 cs.RO、cs.CV

AI总结 该研究提出用语义辐射场（SRF）构建兼具几何真实性与语义可查询性的真实场景模拟器，用于训练评估具身智能体的空间推理能力，还将其应用于果园苹果采摘任务。

Comments Accepted at the IJCAI 2026 Workshop on Spatio-Temporal Reasoning and Learning (STRL), oral presentation

FUSE: Active Functional Affordance Grounding through Adaptive Semantic-Geometric Evidence Acquisition

FUSE：通过自适应语义-几何证据获取实现主动功能可供性接地

Zhou Chen, Sathyanarayanan N. Aakur

专题命中 具身推理 ：embodied agent(abstract)；分类 cs.RO、cs.CV

AI总结 该研究提出主动功能可供性接地任务，构建FUSE框架结合不确定性驱动探索与摊销规划器，基于Habitat基准验证其在非神示接地中性能最优且计算量降低1.33倍。

Comments Under review. 15 Pages. 9 tables, 3 Figures

R4DSG: Relative 4D Scene Graph Memory for Object-Centric Question Answering in Long Egocentric Video

R4DSG：面向长时序自我中心视频中以对象为中心的问答的相对4D场景图记忆

Ke Ma, Yamin Mao, Weiming Li, Shuai Tan, Yijie Zhong, Hao Chen, Haofen Wang, Meng Wang

机构 * Samsung R&D Institute China - Beijing（三星中国研发研究院（北京）） ； Shanghai Jiao Tong University（上海交通大学） 纠错

专题命中 具身推理 ：world model(abstract)；分类 cs.AI、cs.CV

AI总结 该研究提出R4DSG，一种面向长时序自我中心视频的相对4D场景图记忆，在EgoLifeQA数据集的对象相关问答任务中，较EgoRAG-Text取得显著性能提升，可作为可穿戴助手等的实用记忆载体。

Comments 10 pages, 3 figures, ACM Multimedia 2026, egocentric video; 3D scene graph; temporal memory; graph retrieval; object-state reasoning; multimodal question answering