检索结果-内蒙古大学图书馆

Imitating Human Selective Attention Using Dual Policy Network for Scanpath Prediction

学校读者我要写书评

暂无评论

Imitating Human Selective Attention Using Dual Policy Networ...

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Kepei Zhang Ge Tong Xuetao Zhang National Key Laboratory of Human-Machine Hybrid Augmented Intelligence Xi’an China National Engineering Research Center of Visual Information and Applications Xi’an China Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University Xi’an China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Understanding how selective attention influences human gaze behaviors is important for behavioral vision and visual psychology. However, most existing scanpath models ignore the internal sub-stages in visual search and do not imitate human-like selective attention. To bridge this gap, a novel inverse reinforcement learning model with a dual policy network (DPNet) is proposed to accurately predict how humans select and shift their attention at different stages of a task. Additionally, to establish the semantic correlation between objects for state representation modeling, this paper employs Augmented Belief Maps (ABMs). Besides, an Option-Viterbi module is introduced to infer sub-task options of the real human scanpaths, learning the sub-task switching during visual search. Experimental results on widely used datasets for visual search justify the effectiveness of our method in terms of scanpath similarity and sub-task switching.

关键词： Visualization Correlation Inverse problems Semantics Psychology Switches Reinforcement learning Signal processing Predictive models Speech processing

Distributed Multi-robot Source Seeking in Unknown Environments with Unknown Number of Sources

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Chen, Lingpeng Kailas, Siva Deolasee, Srujan Luo, Wenhao Sycara, Katia Kim, Woojun Shenzhen Institute of Artificial Intelligence and Robotics for Society Chinese University of Hong Kong Shenzhen China Robotics Institute Carnegie Mellon University United States School of Interactive Computing Georgia Institute of Technology United States Department of Computer Science University of Illinois Chicago United States

We introduce a novel distributed source seeking framework, DIAS, designed for multi-robot systems in scenarios where the number of sources is unknown and potentially exceeds the number of robots. Traditional robotic source seeking methods typically focused on directing each robot to a specific strong source and may fall short in comprehensively identifying all potential sources. DIAS addresses this gap by introducing a hybrid controller that identifies the presence of sources and then alternates between exploration for data gathering and exploitation for guiding robots to identified sources. It further enhances search efficiency by dividing the environment into Voronoi cells and approximating source density functions based on Gaussian process regression. Additionally, DIAS can be integrated with existing source seeking algorithms. We compare DIAS with existing algorithms, including DoSS and GMES in simulated gas leakage scenarios where the number of sources outnumbers or is equal to the number of robots. The numerical results show that DIAS outperforms the baseline methods in both the efficiency of source identification by the robots and the accuracy of the estimated environmental density function. Copyright © 2025, The Authors. All rights reserved.

关键词： Multipurpose robots

Parameter Efficient Merging for Multimodal Large Language Models with Complementary Parameter Adaptation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zeng, Fanhu Guo, Haiyang Zhu, Fei Shen, Li Tang, Hao State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences China Centre for Artificial Intelligence and Robotics HKISI-CAS Hong Kong School of Cyber Science and Technology Sun Yat-sen University China School of Computer Science Peking University China

Fine-tuning pre-trained models with custom data leads to numerous expert models on specific tasks. Merging models into one universal model to empower multi-task ability refraining from data leakage has gained popularity. With the expansion in data and model size, parameter efficient tuning becomes the common practice for obtaining task-specific models efficiently. However, we observe that existing methods designed for full fine-tuning merging fail under efficient tuning. To address the issues, we analyze from low-rank decomposition and reveal that maintaining direction and compensating for gap between singular values are crucial for efficient model merging. Consequently, we propose CoPA-Merging, a training-free parameter efficient merging method with complementary parameter adaptation. Specifically, we (1) prune parameters and construct scaling coefficients from inter-parameter relation to compensate for performance drop from task interference and (2) perform cross-task normalization to enhance unseen task generalization. We establish a benchmark consisting of diverse multimodal tasks, on which we conduct experiments to certificate the outstanding performance and generalizability of our method. Additional study and extensive analyses further showcase the effectiveness. Copyright © 2025, The Authors. All rights reserved.

关键词： Information leakage

Efficient Backbone Network Construction in Wireless artificial Intelligent Computing Systems

学校读者我要写书评

暂无评论

Tsinghua Science and Technology 2025年第5期30卷 2300-2319页

作者： Ming Sun Xinyu Wu Yi Zhou Jin-Kao Hao Zhang-Hua Fu School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China LERIA Université d'Angers Angers France Shenzhen Institute of Artificial Intelligence and Robotics for Society the Chinese University of Hong Kong Shenzhen China

In wireless artificial intelligent computing systems, the construction of backbone network, which determines the optimum network for a set of given terminal nodes like users, switches, and concentrators, can be naturally formed as the Steiner tree problem. The Steiner tree problem asks for a minimum edge-weighted tree spanning a given set of terminal vertices from a given graph. As a well-known graph problem, many algorithms have been developed for solving this computationally challenging problem in the past decades. However, existing algorithms typically encounter difficulties for solving large instances, i.e., graphs with a high number of vertices and terminals. In this paper, we present a novel partition-and-merge algorithm for effectively handle large-scale graphs. The algorithm breaks the input network into small subgraphs and then merges the subgraphs in a bottom-up manner. In the merging procedure, partial Steiner trees in the subgraphs are also created and optimized by an efficient local optimization. When the merging procedure ends, the algorithm terminates and reports the final solution for the input graph. We evaluated the algorithm on a wide range of benchmark instances, showing that the algorithm outperforms the best-known algorithms on large instances and competes favorably with them on small or middle-sized instances.

关键词： Steiner trees Wireless communication Heuristic algorithms Merging Benchmark testing Cyber-physical systems Search problems Graph theory Partitioning algorithms Optimization

FSVS-Net: A few-shot semi-supervised vessel segmentation network for multiple organs based on feature distillation and bidirectional weighted fusion

学校读者我要写书评

暂无评论

Information Fusion 2025年 123卷

作者： Yang, Yuqun Xu, Jichen Xu, Mengyuan Tang, Xu Wang, Bo Shu, Kechen You, Zheng Institute of Medical Equipment Science and Engineering Huazhong University of Science and Technology Wuhan430074 China National Key Laboratory of Human-Machine Hybrid Augmented Intelligence National Engineering Research Center for Visual Information and Applications and Institute of Artificial Intelligence and Robotics Xi'an Jiaotong University Xi'an710049 China School of Computer Science and Technology Xi'an Jiaotong University Xi'an710049 China School of Artificial Intelligence Xidian University Xi'an710071 China

Accurate 3D vessel mapping is essential for surgical planning and interventional treatments. However, the conventional manual slice-by-slice annotation in CT scans is extremely time-consuming, due to the complexity of vessels: sparse distribution, intricate 3D topology, varying sizes, irregular shapes, and low contrast with the background. To address this problem, we propose a few-shot semi-supervised vessel segmentation network (FSVS-Net) applicable to multiple organs. It can leverage a few annotated slices to segment vessel regions in unannotated slices, enabling efficient semi-supervised processing of the entire CT sequences. Specifically, we propose a feature distillation module for FSVS-Net to enhance vessel-specific semantic representations and suppress irrelevant background features. In addition, we design a bidirectional weighted fusion strategy that propagates information from a few annotated slices to unannotated ones in both opposite directions of the CT sequence, effectively modeling 3D vessel continuity and reducing error accumulation. Extensive experiments on three datasets (hepatic vessels, pulmonary vessels, and renal arteries) demonstrate that FSVS-Net achieves state-of-the-art performance in few-shot vessel segmentation task, significantly outperforming existing methods. We collected and annotated three vessel datasets, including clinical data from Tsinghua Changgung Hospital and public sources (e.g., MSD08), for this study. In practice, it reduces the average annotation time from 2 h to 0.5 h per volume, improving efficiency by 4×. We release three organ-specific vessel datasets and the implementation code at: https://***/YqunYang/FSVS-Net. © 2025

关键词： Semi-supervised learning

KISS-SLAM: A Simple, Robust, and Accurate 3D LiDAR SLAM System With Enhanced Generalization Capabilities

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Guadagnino, Tiziano Mersch, Benedikt Gupta, Saurabh Vizzo, Ignacio Grisetti, Giorgio Stachniss, Cyrill Center for Robotics University of Bonn Germany Sapienza University of Rome Italy Department of Engineering Science the University of Oxford United Kingdom Lamarr Institute for Machine Learning and Artificial Intelligence Germany

Robust and accurate localization and mapping of an environment using laser scanners, so-called LiDAR SLAM, is essential to many robotic applications. Early 3D LiDAR SLAM methods often exploited additional information from IMU or GNSS sensors to enhance localization accuracy and mitigate drift. Later, advanced systems further improved the estimation at the cost of a higher runtime and complexity. This paper explores the limits of what can be achieved with a LiDAR-only SLAM approach while following the "Keep It Small and Simple" (KISS) principle. By leveraging this minimalistic design principle, our system, KISS-SLAM, archives state-of-the-art performances in pose accuracy while requiring little to no parameter tuning for deployment across diverse environments, sensors, and motion profiles. We follow best practices in graph-based SLAM and build upon LiDAR odometry to compute the relative motion between scans and construct local maps of the environment. To correct drift, we match local maps and optimize the trajectory in a pose graph optimization step. The experimental results demonstrate that this design achieves competitive performance while reducing complexity and reliance on additional sensor modalities. By prioritizing simplicity, this work provides a new strong baseline for LiDAR-only SLAM and a high-performing starting point for future research. Further, our pipeline builds consistent maps that can be used directly for further downstream tasks like navigation. Our open-source system operates faster than the sensor frame rate in all presented datasets and is designed for real-world scenarios. Copyright © 2025, The Authors. All rights reserved.

关键词： Robot applications

Physics-Embedded Motion Planning With Contact Handling for Continuum Surgical Robots

学校读者我要写书评

暂无评论

IEEE robotics and Automation Letters 2025年第7期10卷 7031-7038页

作者： Yixiong Du Yi Xiong Mingcong Chen Guokai Zhang Hao Wang Lu Liu Hongbin Liu Zhongkai Zhang Department of Biomedical Engineering City University of Hong Kong Hong Kong Centre for Artificial Intelligence and Robotics Hong Kong Institute of Science & Innovation Chinese Academy of Sciences Hong Kong Interdisciplinary Science Hub (ISH) Hypercycle Venture China

Motion planning for Continuum Surgical Robots (CSRs) faces significant challenges during minimally invasive surgery (MIS) when operating within highly constrained anatomical workspaces. Although traditional approaches aim to prevent tissue damage by finding collision-free paths, robot-tissue interactions are often unavoidable in practical surgical scenarios. To address this, we present a physics-embedded motion planning framework where dynamic interactions between CSRs and tissues are modeled using real-time finite element method (FEM). Our framework integrates computed contact forces as safety constraints into an adaptive sampling-based motion planning algorithm enhanced by Gaussian mixture model (GMM) optimization for efficient exploration. We validate the proposed framework in anatomically accurate scenarios involving the brain ventricles and the tracheobronchial tree, using two types of CSRs: the concentric tube robot and the cable-driven continuum robot. The results demonstrate that our method outperforms the traditional collision-free strategy in terms of path quality and efficiency. Notably, our method provides feasible solutions in scenarios where collision-free paths are unattainable. Lastly, the applicability of the proposed framework is validated in a real-world experiment.

关键词： Robots Planning Collision avoidance Finite element analysis Mathematical models Real-time systems Computational modeling Navigation Safety Medical robotics

Florenz: Scaling Laws for Systematic Generalization in Vision-Language Models

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Spravil, Julian Houben, Sebastian Behnke, Sven Fraunhofer IAIS Germany University of Applied Sciences Bonn-Rhein-Sieg Bonn Germany University of Bonn Computer Science Institute VI Center for Robotics Germany Lamarr Institute for Machine Learning and Artificial Intelligence Germany

Cross-lingual transfer enables vision-language models (VLMs) to perform vision tasks in various languages with training data only in one language. Current approaches rely on large pre-trained multilingual language models. However, they face the curse of multilinguality, sacrificing downstream task performance for multilingual capabilities, struggling with lexical ambiguities, and falling behind recent advances. In this work, we study the scaling laws of systematic generalization with monolingual VLMs for multilingual tasks, focusing on the impact of model size and seen training samples. We propose Florenz, a monolingual encoder-decoder VLM with 0.4B to 11.2B parameters combining the pre-trained VLM Florence-2 and the large language model Gemma-2. Florenz is trained with varying compute budgets on a synthetic dataset that features intentionally incomplete language coverage for image captioning, thus, testing generalization from the fully covered translation task. We show that not only does indirectly learning unseen task-language pairs adhere to a scaling law, but also that with our data generation pipeline and the proposed Florenz model family, image captioning abilities can emerge in a specific language even when only data for the translation task is available. Fine-tuning on a mix of downstream datasets yields competitive performance and demonstrates promising scaling trends in multimodal machine translation (Multi30K, CoMMuTE), lexical disambiguation (CoMMuTE), and image captioning (Multi30K, XM3600, COCO Karpathy). Copyright © 2025, The Authors. All rights reserved.

关键词： Budget control

Practical Continual Forgetting for Pre-trained Vision Models

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhao, Hongbo Zhu, Fei Ni, Bolin Zhu, Feng Meng, Gaofeng Zhang, Zhaoxiang State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China University of Chinese Academy of Sciences Beijing100190 China Centre for Artificial Intelligence and Robotics Hong Kong Institute of Science & Innovation Chinese Academy of Sciences 999077 Hong Kong SenseTime Research Beijing100084 China

For privacy and security concerns, the need to erase unwanted information from pre-trained vision models is becoming evident nowadays. In real-world scenarios, erasure requests originate at any time from both users and model owners, and these requests usually form a sequence. Therefore, under such a setting, selective information is expected to be continuously removed from a pre-trained model while maintaining the rest. We define this problem as continual forgetting and identify three key challenges. (i) For unwanted knowledge, efficient and effective deleting is crucial. (ii) For remaining knowledge, the impact brought by the forgetting procedure should be minimal. (iii) In real-world scenarios, the training samples may be scarce or partially missing during the process of forgetting. To address them, we first propose Group Sparse LoRA (GS-LoRA). Specifically, towards (i), we introduce LoRA modules to fine-tune the FFN layers in Transformer blocks for each forgetting task independently, and towards (ii), a simple group sparse regularization is adopted, enabling automatic selection of specific LoRA groups and zeroing out the others. To further extend GS-LoRA to more practical scenarios, we incorporate prototype information as additional supervision and introduce a more practical approach, GS-LoRA++. For each forgotten class, we move the logits away from its original prototype. For the remaining classes, we pull the logits closer to their respective prototypes. We conduct extensive experiments on face recognition, object detection and image classification and demonstrate that our method manages to forget specific classes with minimal impact on other classes. Codes have been released on https://***/bjzhb666/GS-LoRA. Copyright © 2025, The Authors. All rights reserved.

关键词： Face recognition