FlowCorrect is a modular interactive imitation learning method that allows real‑time adaptation of generative flow‑matching manipulation policies using sparse, relative human corrections. During task execution, a human provides brief corrective pose nudges through a lightweight VR interface, and FlowCorrect locally adapts the policy without retraining the backbone, maintaining performance on previously learned scenarios. Experiments on a real‑world robot across four tabletop tasks show that, with a low correction budget, FlowCorrect achieves an 80% success rate on previously failed cases while preserving performance on solved scenarios.
By Edgar Welte, Yitian Shi, Rosa Wolf, Maximillian Gilles, Rania Rayyes
The paper presents a method for reconstructing object geometry under occlusion by combining generative shape priors with contact-based constraints. Generative models provide plausible guesses for unseen parts, while contact information from videos or physical interactions supplies sparse boundary constraints. The authors integrate these cues through a contact-guided 3D generation framework, demonstrating improved reconstruction on synthetic and real-world datasets compared to baseline approaches.
By Minghan Zhu, Zhiyi Wang, Qihang Sun, Maani Ghaffari, Michael Posa
The paper introduces VLAct, a Vision‑Language‑Action model that focuses on representation‑centric continued pre‑training rather than merely scaling robot data. VLAct is trained on diverse, multi‑embodiment robot data and preserves a broad VLM prior while encouraging shared action semantics across embodiments. Experiments across simulation, real‑world, and unseen‑embodiment settings show that VLAct consistently outperforms existing industrial VLA systems, achieving high success rates with only a modest compute budget and open‑source data.
By Senqiao Yang, Chengyao Wang, Yuxin Chen, Zixuan Wang, Longxiang Tang, Haokun Gui, Jinhui Ye, Changsheng Lu, Xiaoyang Wu, Mingkang Zhu, Pengguang Chen, Shu Liu, Zhuotao Tian, Hengshuang Zhao, Bei Yu, Jiaya Jia
The paper proposes a label‑free 3D perception framework where roadside units (RSUs) act as unsupervised teachers for self‑driving cars. RSUs learn local 3D detectors from unlabeled data and broadcast predictions to passing vehicles, which use these as pseudo‑labels to train an ego‑centric detector. In a CARLA simulation, the method achieves 82.3% AP for vehicle detection, approaching a fully supervised upper bound of 94.4%, and demonstrates scalability and complementarity with existing ego‑centric approaches.
By Zhen Xu, Jinsu Yoo, Cristian Bautista, Zanming Huang, Tai-Yu Pan, Zhenzhen Liu, Katie Z Luo, Mark Campbell, Bharath Hariharan, Wei-Lun Chao
RegCL is a non‑replay continual learning framework that adapts the Segment Anything Model (SAM) for visual grounding across evolving multi‑sensorial media domains. It consolidates domain‑specific segmentation knowledge into a single lightweight SAM adapter by incrementally merging LoRA‑style AugModules and preserving compact historical feature statistics. Experiments on five heterogeneous datasets demonstrate that RegCL retains performance while adapting to new domains, outperforming other non‑replay continual learning and merging baselines.
By Yuan-Chen Shu, Zhiwei Lin, Xiaoyu Zhou, Yongtao Wang
Aero Hand Open is a tendon‑driven, anthropomorphic hand designed for affordable, simulation‑ready dexterous manipulation. The release includes a realistic simulation model of the cable transmission, an identified actuation map that links motor commands to joint motions (including thumb coupling), and a reinforcement‑learning package that trains policies entirely in simulation. These components enable policies to run on the physical hand without fine‑tuning or state estimation.
By Nan Wang, Mohit Yadav, Jonathan Wulff, Aidan Rosenbaum, Kezhou Chen, Yuvan Sharma, Xu Dong, Yiwei Tao
Multimodal Large Language Models (MLLMs) process speech and text jointly, yet whether they exploit prosodic cues for pragmatic inference or rely on surface acoustic patterns has received little system...
The paper introduces a unified conditional-flow framework that integrates text-driven motion generation, semantic editing, and intra-structural retargeting into a single rectified-flow model. By treating editing as a change in semantic condition and retargeting as a change in skeletal condition, the approach eliminates fragmented pipelines and allows a single model to perform generation, zero‑shot editing, and zero‑shot retargeting on articulated 3D motion data. Experiments on SnapMoGen and a Mixamo subset demonstrate that the model can handle all three tasks without task‑specific fine‑tuning, preserving both motion semantics and skeletal structure.
By Junlin Li, Xinhao Song, Siqi Wang, Haibin Huang, Yili Zhao
DeflectBench is a new benchmark that evaluates how large language models (LLMs) generate rhetorical fallacies when prompted. The study tests 23,990 generations from four leading models using three deflection strategies (whataboutism, ad hominem, red herring), seven prompt framings, and 80 claims across four controversy levels. Results show that refusal to produce fallacies depends mainly on request structure, with prompt framing and fallacy type dramatically affecting compliance rates.
By Art Kanke
CLAP is a cross-embodiment framework for action‑conditioned video generation that can be trained on diverse internet‑scale videos from both humans and robots. It reconciles different action spaces—end‑effector poses, language instructions, and latent actions—using a curriculum that first learns physics priors from unlabeled video and then grounds them in real‑world action spaces for zero‑shot deployment. The resulting models match or exceed state‑of‑the‑art single‑embodiment models in challenging environments and support few‑shot adaptation across a wide range of robot morphologies.
By Kechen Liu, Ola Shorinwa
The paper applies the AICON reactive gradient‑descent framework, originally designed for robotic manipulation, to the Tower of London cognitive test. AICON, without any lookahead planning or human cognition knowledge, reproduces the fine‑grained difficulty ordering of 24 problems better than structural task parameters and generalizes to held‑out problems. It outperforms a planning baseline for groups with reduced planning capacity (e.g., Parkinson’s patients) while the baseline better captures healthy controls, indicating that reduced planning capacity shifts human behavior toward a reactive mode similar to AICON’s failure patterns.
By Michael Migacev, Vito Mengers, Antonia K\"ongeter, Oliver Brock
The paper investigates how the structural entanglement of facts within a large language model’s knowledge base influences whether those facts leak after unlearning. Using two unlearning algorithms (WHP and GA+KL) across fictional and real-world datasets and multiple model sizes, the authors find that highly entangled facts are more likely to be recalled before unlearning, but the relationship changes—WHP weakens it while GA+KL reverses it. By directly manipulating entanglement scores and observing corresponding recall changes, they demonstrate a causal link and develop a predictive tool to audit prompts for potential leakage.
By Aakriti Shah, Yifan Hu, Thai Le
DINOcular is a self‑supervised framework that learns joint visuospatial representations from RGB‑D observations. It fuses depth‑derived geometric priors with a visual backbone using inter‑patch and intra‑patch fusion, allowing the model to encode both appearance and spatial structure efficiently. The resulting representation improves 3D awareness on multiple geometry benchmarks while staying competitive on standard RGB‑D semantic segmentation tasks.
By Farkhat Almukhamedov, Sami Azirar, Hermann Blum
EgoSurg is a framework that reconstructs dynamic operating room scenes from sparse wall‑mounted stereo video and renders arbitrary, role‑specific egocentric views without instrumenting personnel. It builds a per‑timestamp 3D Gaussian Splatting representation using scale‑aware stereo depth and refines it with an image‑conditioned diffusion model to correct artifacts from limited camera coverage, crowding, and occlusion. Evaluations on real robotic pulmonology procedures and simulated sessions show consistent near‑field reconstruction fidelity (PSNR 26.8 dB, SSIM 0.895) and synthesized egocentric view quality (PSNR 17.8 dB, SSIM 0.766) across workflow phases and hospital sites, with case studies demonstrating applications such as sterile field violation adjudication, role‑specific replay, and counterfactual personnel positioning.
By Han Zhang, Lalithkumar Seenivasan, Jose L. Porras, Roger D. Soberanis-Mukul, Hao Ding, Hongchao Shu, Benjamin D. Killeen, Ankita Ghosh, Lonny Yarmus, Jeffrey K. Jopling, Masaru Ishii, Angela C. Argento, Mathias Unberath
The paper introduces Disjoint Parameter Training (DPT), a framework that addresses Skill Conflict—where shared encoder parameters hinder separate tasks of motion prediction and safety planning—by training tasks on distinct parameter subsets before merging. DPT employs sparse merging to integrate only the most influential parameters, reducing interference and enhancing representational capacity. Experiments on JRDB and JTA benchmarks show that DPT outperforms existing unified models, demonstrating its effectiveness for safe, resource‑efficient robot navigation.
By Taewon Seo, Seonae Jeon, Giwon Lee, Kuk-Jin Yoon, Daehee Park
The paper introduces Residual Reward Models (RRM) to enhance preference‑based reinforcement learning (PbRL) in robotics. RRMs decompose the true reward into a prior component—such as a heuristic, language‑generated, or IRL‑derived reward—and a learned residual that is trained with human preferences. Experiments on Meta‑World, DM‑Control, and a physical Franka Panda robot show that RRMs markedly improve sample efficiency and accelerate policy learning compared to standard PbRL methods.
By Chenyang Cao, Miguel Rogel-Garc\'ia, Mohamed Nabail, Xueqian Wang, Nicholas Rhinehart
The paper tackles the problem of calibrating a stationary camera’s yaw, pitch, and roll using a drone’s GPS trajectory, which suffers from altitude bias, time offset, and discrete sampling. It formulates a parameter estimation problem that jointly estimates the GPS altitude bias, time offset, and camera orientation biases, and proposes a maximum likelihood estimator based on Iterated Least Squares to handle the asynchronous, discrete GPS data. Simulation results show the estimator achieves accuracy close to the Cramér–Rao Lower Bound, with a recommended drone trajectory yielding calibration errors within 14% of the measurement error standard deviation.
By R. Yang, Y. Bar-Shalom, H. A. J. Huang
The paper introduces sLoTh, a parameter‑efficient continual learning framework for sparse event‑based vision transformers. sLoTh freezes the backbone and limits plasticity to low‑rank attention updates (seLoRA) and shared neuronal threshold modulation, updating less than 1% of parameters without replay buffers. Experiments on CIFAR‑100, Tiny‑ImageNet, ImageNet‑100, and ImageNet‑R show competitive rehearsal‑free performance across up to 100 tasks while achieving roughly 6.5× lower energy consumption than dense vision transformers.
By Vaishnavi Nagabhushana, Kartikay Agrawal, Ayon Borthakur
MineExplorer is a benchmark designed to assess the open‑world exploration abilities of multimodal large language models (MLLMs) in Minecraft. It filters out tasks that rely heavily on Minecraft‑specific knowledge, organizes tasks into ReAct‑style capabilities, and composes atomic tasks into implicit multi‑hop challenges. A multi‑agent synthesis workflow creates reliable task graphs, sandbox scenes, and rule‑based milestone evaluators, and human evaluation confirms its superiority over a single‑agent baseline. Experiments show that while advanced MLLMs can handle many single‑hop tasks, they struggle with longer trajectories that require coordinating hidden prerequisites, and larger models or different thinking modes do not consistently improve performance.
By Tianjie Ju, Yueqing Sun, Zheng Wu, Wei Zhang, Yaqi Huo, Xi Su, Qi Gu, Xunliang Cai, Gongshen Liu, Zhuosheng Zhang
The paper introduces a new approach to safety in contextual bandits with continuous actions, focusing on high‑probability constraints on the realized cost rather than expected cost. It presents the High‑Probability Constrained UCB algorithm, which balances reward exploration with conservative safety estimation, and provides theoretical regret guarantees for linear models and extensions to general function classes. Experiments demonstrate that this realized‑cost safety framework significantly reduces safety violations compared to expected‑cost constrained methods.
By Spyros Dragazis, Aldo Pacchiano