检索结果-内蒙古大学图书馆

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Lee, Donghoon Luu, Tung M. Lee, Younghwan Yoo, Chang D. Robotics Program KAIST Daejeon Korea Republic of Electrical Engineering KAIST Daejeon Korea Republic of

ISBN: (纸本)9798350368741

Recent research highlights the potential of multimodal foundation models in tackling complex decision-making challenges. However, their large parameters make real-world deployment resource-intensive and often impractical for constrained systems. Reinforcement learning (RL) shows promise for task-specific agents but suffers from high sample complexity, limiting practical applications. To address these challenges, we introduce LVLM to Policy (LVLM2P), a novel framework that distills knowledge from large vision-language models (LVLM) into more efficient RL agents. Our approach leverages the LVLM as a teacher, providing instructional actions based on trajectories collected by the RL agent, which helps reduce less meaningful exploration in the early stages of learning, thereby significantly accelerating the agent's learning progress. Additionally, by leveraging the LVLM to suggest actions directly from visual observations, we eliminate the need for manual textual descriptors of the environment, enhancing applicability across diverse tasks. Experiments show that LVLM2P significantly enhances the sample efficiency of baseline RL algorithms. The code is available at https://***/i22024/LVLM2P. © 2025 IEEE.

关键词： Knowledge Distillation Reinforcement Learning Vision Language Model

来源：评论

学校读者我要写书评

暂无评论

Reward Generation via Large Vision-Language Model in Offline Reinforcement Learning

Reward Generation via Large Vision-Language Model in Offline...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Lee, Younghwan Luu, Tung M. Lee, Donghoon Yoo, Chang D. Electrical Engineering KAIST Daejeon Korea Republic of Robotics Program KAIST Daejeon Korea Republic of

ISBN: (纸本)9798350368741

In offline reinforcement learning (RL), learning from fixed datasets presents a promising solution for domains where real-time interaction with the environment is expensive or risky. However, designing dense reward signals for offline dataset requires significant human effort and domain expertise. Reinforcement learning with human feedback (RLHF) has emerged as an alternative, but it remains costly due to the human-in-the-loop process, prompting interest in automated reward generation models. To address this, we propose Reward Generation via Large Vision-Language Models (RG-VLM), which leverages the reasoning capabilities of LVLMs to generate rewards from offline data without human involvement. RG-VLM improves generalization in long-horizon tasks and can be seamlessly integrated with the sparse reward signals to enhance task performance, demonstrating its potential as an auxiliary reward signal. © 2025 IEEE.

关键词： Large Vision-Language Model Offline Reinforcement Learning Reward Labeling

来源：评论

学校读者我要写书评

暂无评论

Designing an End-to-End UAV System for Insulator Inspection Under Transmission Tower Environments

Designing an End-to-End UAV System for Insulator Inspection ...

引用

2024 International Conference on Electronics, Information, and Communication, ICEIC 2024

作者： Jeong, Jinyeong Park, Seongsu Lee, Sanghoon Youn, Donghyeon Kim, Min Jun School of Electrical Engineering KAIST Daejeon Korea Republic of KAIST Robotics Program Daejeon Korea Republic of KEPCO Research Institute Korea Electric Power Corporation Daejeon Korea Republic of

ISBN: (纸本)9798350371888

This paper proposes an unmanned aerial vehicle (UAV) equipped with a rigid contact module and two supplementary forward-direction motors, designed for the contact-based inspection of porcelain insulators. During the free-flight, this aerial robot is controlled by a geometric tracking controller. When approaching and contacting the insulators, the impedance controller generates enough force to sustain contact by using two forward-direction motors, while the geometric tracking controller maintains its orientation. This platform is validated through simulations, demonstrating its effectiveness in performing contact-based inspection tasks. © 2024 IEEE.

关键词： Inspection

来源：评论

学校读者我要写书评

暂无评论

ED-NERF: EFFICIENT TEXT-GUIDED EDITING OF 3D SCENE WITH LATENT SPACE NERF 12

ED-NERF: EFFICIENT TEXT-GUIDED EDITING OF 3D SCENE WITH LATE...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Park, Jangho Kwon, Gihyun Ye, Jong Chul Kim Jaechul Graduate School of AI Korea Republic of Robotics Program Korea Republic of Department of Bio and Brain Engineering KAIST Korea Republic of

Recently, there has been a significant advancement in text-to-image diffusion models, leading to groundbreaking performance in 2D image generation. These advancements have been extended to 3D models, enabling the generation of novel 3D objects from textual descriptions. This has evolved into NeRF editing methods, which allow the manipulation of existing 3D objects through textual conditioning. However, existing NeRF editing techniques have faced limitations in their performance due to slow training speeds and the use of loss functions that do not adequately consider editing. To address this, here we present a novel 3D NeRF editing approach dubbed ED-NeRF by successfully embedding real-world scenes into the latent space of the latent diffusion model (LDM) through a unique refinement layer. This approach enables us to obtain a NeRF backbone that is not only faster but also more amenable to editing compared to traditional image space NeRF editing. Furthermore, we propose an improved loss function tailored for editing by migrating the delta denoising score (DDS) distillation loss, originally used in 2D image editing to the three-dimensional domain. This novel loss function surpasses the well-known score distillation sampling (SDS) loss in terms of suitability for editing purposes. Our experimental results demonstrate that ED-NeRF achieves faster editing speed while producing improved output quality compared to state-of-the-art 3D editing models. Code and rendering results are available at our project page. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

Sample Efficient Reinforcement Learning via Large Vision Language Model Distillation

Sample Efficient Reinforcement Learning via Large Vision Lan...

引用

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

作者： Donghoon Lee Tung M. Luu Younghwan Lee Chang D. Yoo Robotics Program KAIST Daejeon South Korea Electrical Engineering KAIST Daejeon South Korea

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Recent research highlights the potential of multi-modal foundation models in tackling complex decision-making challenges. However, their large parameters make real-world deployment resource-intensive and often impractical for constrained systems. Reinforcement learning (RL) shows promise for task-specific agents but suffers from high sample complexity, limiting practical applications. To address these challenges, we introduce LVLM to Policy (LVLM2P), a novel framework that distills knowledge from large vision-language models (LVLM) into more efficient RL agents. Our approach leverages the LVLM as a teacher, providing instructional actions based on trajectories collected by the RL agent, which helps reduce less meaningful exploration in the early stages of learning, thereby significantly accelerating the agent’s learning progress. Additionally, by leveraging the LVLM to suggest actions directly from visual observations, we eliminate the need for manual textual descriptors of the environment, enhancing applicability across diverse tasks. Experiments show that LVLM2P significantly enhances the sample efficiency of baseline RL algorithms. The code is available at https://***/i22024/LVLM2P

关键词： Visualization Limiting Foundation models Decision making Signal processing algorithms Reinforcement learning Cognition Trajectory Complexity theory Speech processing

来源：评论

学校读者我要写书评

暂无评论

Reward Generation via Large Vision-Language Model in Offline Reinforcement Learning

Reward Generation via Large Vision-Language Model in Offline...

引用

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

作者： Younghwan Lee Tung M. Luu Donghoon Lee Chang D. Yoo Electrical Engineering KAIST Daejeon South Korea Robotics Program KAIST Daejeon South Korea

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

关键词： Reinforcement learning Signal processing Real-time systems Human in the loop Data models Cognition Acoustics Labeling Speech processing

来源：评论

学校读者我要写书评

暂无评论

Cognitive TransFuser: Semantics-Guided Transformer-Based Sensor Fusion for Improved Waypoint Prediction 11th

Cognitive TransFuser: Semantics-Guided Transformer-Based Sen...

引用

11th International Conference on Robot Intelligence Technology and Applications, RiTA 2023

作者： Choi, Hwan-Soo Jeong, Jongoh Hoo Cho, Young Yoon, Kuk-Jin Kim, Jong-Hwan Robot Intelligence Technology Lab. School of Engineering KAIST Daejeon Korea Republic of dSPECTER Seoul Korea Republic of Visual Intelligence Lab. Robotics Program and Dept. of Mechanical Engineering KAIST Daejeon Korea Republic of

ISBN: (纸本)9783031706868

Sensor fusion approaches for intelligent self-driving agents remain key to driving scene understanding given visual global contexts acquired from input sensors. Specifically, for the local waypoint prediction task, single-modality networks are still limited by strong dependency on the sensitivity of the input sensor, and thus recent works therefore promote the use of multiple sensors in fusion in feature level in practice. While it is well known that multiple data modalities encourage mutual contextual exchange, it requires global 3D scene understanding in real-time with minimal computation upon deployment to practical driving scenarios, thereby placing greater significance on the training strategy given a limited number of practically usable sensors. In this light, we exploit carefully selected auxiliary tasks that are highly correlated with the target task of interest (e.g., traffic light recognition and semantic segmentation) by fusing auxiliary task features and also using auxiliary heads for waypoint prediction based on imitation learning. Our RGB-LIDAR-based multi-task feature fusion network, coined Cognitive TransFuser, augments and exceeds the baseline network by a significant margin for safer and more complete road navigation in the CARLA simulator. We validate the proposed network on the Town05 Short and Town05 Long Benchmark through extensive experiments, achieving up to 44.2 FPS real-time inference time. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Conditional Latent ODEs for Motion Prediction in Autonomous Driving

arXiv

引用

arXiv 2024年

作者： Khang, Truong G. Kim, Yongjae Finazzi, Andrea KAIST CS Korea Republic of KAIST ME Korea Republic of The Robotics Program - KAIST EE Korea Republic of

This paper addresses imitation learning for motion prediction problem in autonomous driving, especially in multi-agent setting. Different from previous methods based on GAN, we present the conditional latent ordinary differential equation (cLODE) to leverage both the generative strength of conditional VAE and the continuous representation of neural ODE. Our network architecture is inspired from the Latent-ODE model. The experiment shows that our method outperform the baseline methods in the simulation of multi-agent driving and is very efficient in term of GPU memory consumption. Our code and docker image are publicly available: https://***/TruongKhang/cLODE;https://***/r/kim4375731/clode. © 2024, CC BY.

关键词： Ordinary differential equations

来源：评论

学校读者我要写书评

暂无评论

RelMobNet: End-to-End Relative Camera Pose Estimation Using a Robust Two-Stage Training 17th

RelMobNet: End-to-End Relative Camera Pose Estimation Using ...

引用

17th European Conference on Computer Vision, ECCV 2022

作者： Rajendran, Praveen Kumar Mishra, Sumit Vecchietti, Luiz Felipe Har, Dongsoo Division of Future Vehicle KAIST Daejeon Korea Republic of The Robotics Program KAIST Daejeon Korea Republic of Data Science Group Institute for Basic Science Daejeon Korea Republic of The CCS Graduate School of Mobility KAIST Daejeon Korea Republic of

ISBN: (纸本)9783031250743

Relative camera pose estimation, i.e. estimating the translation and rotation vectors using a pair of images taken in different locations, is an important part of systems in augmented reality and robotics. In this paper, we present an end-to-end relative camera pose estimation network using a siamese architecture that is independent of camera parameters. The network is trained using the Cambridge Landmarks data with four individual scene datasets and a dataset combining the four scenes. To improve generalization, we propose a novel two-stage training that alleviates the need of a hyperparameter to balance the translation and rotation loss scale. The proposed method is compared with one-stage training CNN-based methods such as RPNet and RCPNet and demonstrate that the proposed model improves translation vector estimation by 16.11%, 28.88%, and 52.27% on the Kings College, Old Hospital, and St Marys Church scenes, respectively. For proving texture invariance, we investigate the generalization of the proposed method augmenting the datasets to different scene styles, as ablation studies, using generative adversarial networks. Also, we present a qualitative assessment of epipolar lines of our network predictions and ground truth poses. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Augmented reality

来源：评论

学校读者我要写书评

暂无评论

Development of a Collaborative Robotic Arm-based Bimanual Haptic Display

arXiv

引用

arXiv 2024年

作者： Lee, Joong-Ku Kim, DongHyeon Park, Seong-Su Lee, Jiye Ryu, Jee-Hwan Department of Civil and Environmental Engineering KAIST Daejeon Korea Republic of Robotics Program KAIST Daejeon Korea Republic of

This paper presents a bimanual haptic display based on collaborative robot arms. We address the limitations of existing robot arm-based haptic displays by optimizing the setup configuration and implementing inertia/friction compensation techniques. The optimized setup configuration maximizes workspace coverage, dexterity, and haptic feedback capability while ensuring collision safety. Inertia/friction compensation significantly improve transparency and reduce user fatigue, leading to a more seamless and transparent interaction. The effectiveness of our system is demonstrated in various applications, including bimanual bilateral teleoperation in both real and simulated environments. This research contributes to the advancement of haptic technology by presenting a practical and effective solution for creating high-performance bimanual haptic displays using collaborative robot arms. © 2024, CC BY.

关键词： Collaborative robots

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：