检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Kim, Heeseung Lee, Che Hyun Park, Sangkwon Yeom, Jiheum Park, Nohil Yu, Sangwon Yoon, Sungroh Department of Electrical and Computer Engineering Seoul National University Korea Republic of AIIS ASRI INMC ISRC and IPAI Seoul National University Korea Republic of

Recent advancements in multi-turn voice interaction models have improved user-model communication. However, while closed-source models effectively retain and recall past utterances, whether open-source models share this ability remains unexplored. To fill this gap, we systematically evaluate how well open-source interaction models utilize past utterances using ContextDialog, a benchmark we proposed for this purpose. Our findings show that speech-based models have more difficulty than text-based ones, especially when recalling information conveyed in speech, and even with retrieval-augmented generation, models still struggle with questions about past utterances. These insights highlight key limitations in open-source models and suggest ways to improve memory retention and retrieval robustness.1 © 2025, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Unsupervised 3D Part Decomposition via Leveraged Gaussian Splatting

Unsupervised 3D Part Decomposition via Leveraged Gaussian Sp...

引用

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

作者： Jae Goo Choy Geonho Cha Hogun Kee Songhwai Oh Sequor Robotics Seoul Korea NAVER Cloud Seongnam Korea Department of Electrical and Computer Engineering and ASRI Seoul National University Seoul Korea

ISBN: (数字)9798350377705

ISBN: (纸本)9798350377712

We propose a novel unsupervised method for motion-based 3D part decomposition of articulated objects using a single monocular video of a dynamic scene. In contrast to existing unsupervised methods relying on optical flow or tracking techniques, our approach addresses this problem without additional information by leveraging Gaussian splatting techniques. We generate a series of Gaussians from a monocular video and analyze their relationships to decompose the dynamic scene into motion-based parts. To decompose dynamic scenes consisting of articulated objects, we design an articulated deformation field suitable for the movement of articulated objects. And to effectively understand the relationships of Gaussians of different shapes, we propose a 3D reconstruction loss using 3D occupied voxel maps generated from the Gaussians. Experimental results demonstrate that our method outperforms existing approaches in terms of 3D part decomposition for articulated objects. More demos and code are available at https://***/artnerf/.

关键词： Training Geometry Three-dimensional displays Codes Tracking Shape Deformation Dynamics Optical flow Intelligent robots

来源：评论

学校读者我要写书评

暂无评论

Fully Distributed EV Charging Scheduling for Load Flattening in V2G Systems

Fully Distributed EV Charging Scheduling for Load Flattening...

引用

IEEE Conference on Decision and Control

作者： Jinwook Heo Soojeong Hyeon Hyungbo Shim Jinsung Kim Department of Electrical and Computer Engineering ASRI Seoul National University Seoul Korea Vehicle SW Division Hyundai Motor Company Hwaseong Korea

ISBN: (数字)9798350316339

ISBN: (纸本)9798350316346

This paper presents a fully distributed algorithm for scheduling electric vehicle (EV) charging and discharging to flatten the total load of the grid, while considering constraints on grid transmission capacity. As a fully distributed solution, the proposed algorithm operates without the need for a central unit. Instead, each agent only communicates a single dual variable with its neighboring agents based on a communication graph, and thus no private information is shared. In particular, the algorithm does not rely on initial conditions, ensuring robustness in online changes of operational conditions. Simulation results verify the effectiveness of the proposed algorithm.

关键词： Vehicle-to-grid Privacy Scheduling algorithms Heuristic algorithms Simulation Scheduling Electric vehicle charging Robustness Vehicle dynamics Distributed algorithms

来源：评论

学校读者我要写书评

暂无评论

Conversion of Controllers to have Integer State Matrix for Encrypted Control: Non-Minimal Order Approach

Conversion of Controllers to have Integer State Matrix for E...

引用

IEEE Conference on Decision and Control

作者： Joowon Lee Donggil Lee Seungbeom Lee Junsoo Kim Hyungbo Shim Department of Electrical and Computer Engineering ASRI Seoul National University Korea Department of Electrical and Information Engineering Seoul National University of Science and Technology Korea

To implement an encrypted dynamic controller based on homomorphic encryption that operates for an infinite time horizon, it is essential for every component of the controller's state matrix to be an integer. In this paper, we tackle the challenge of converting a pre-designed controller into a new one with an integer state matrix, while preserving its control performance. This enables encrypted dynamic systems to be realized without re-encryption and approximation of control parameters. To achieve this, we allow the order of the controller to be increased so that the resulting closed-loop system becomes a non-minimal realization of the original closed-loop, without losing internal stability. Two approaches are proposed to design such controller with an integer state matrix. The first approach is to design the new controller as an estimator of the original closed-loop system, and the conditions on the estimator gain are derived. Our second approach is to formulate a problem of finding certain polynomials, whose solution leads to the design of the new controller. In a special case when the numerator of the plant transfer function is a constant, we provide a constructive method to obtain such solution.

关键词：

来源：评论

学校读者我要写书评

暂无评论

NanoVoice: Efficient Speaker-Adaptive Text-to-Speech for Multiple Speakers

NanoVoice: Efficient Speaker-Adaptive Text-to-Speech for Mul...

引用

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

作者： Nohil Park Heeseung Kim Che Hyun Lee Jooyoung Choi Jiheum Yeom Sungroh Yoon Electrical and Computer Engineering Seoul National University Seoul Republic of Korea ECE AIIS ASRI INMC ISRC and IPAI Seoul National University Seoul Republic of Korea

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

We present NanoVoice, a personalized text-to-speech model that efficiently constructs voice adapters for multiple speakers simultaneously. NanoVoice introduces a batch-wise speaker adaptation technique capable of fine-tuning multiple references in parallel, significantly reducing training time. Beyond building separate adapters for each speaker, we also propose a parameter sharing technique that reduces the number of parameters used for speaker adaptation. By incorporating a novel trainable scale matrix, NanoVoice mitigates potential performance degradation during parameter sharing. NanoVoice achieves performance comparable to the baselines, while training 4 times faster and using 45 percent fewer parameters for speaker adaptation with 40 reference voices. Extensive ablation studies and analysis further validate the efficiency of our model.

关键词： Training Degradation Adaptation models Analytical models Buildings Signal processing Acoustics Text to speech Speech processing Commercialization

来源：评论

学校读者我要写书评

暂无评论

REMOVING UNDESIRABLE FEATURE CONTRIBUTIONS USING OUT-OF-DISTRIBUTION DATA 9

REMOVING UNDESIRABLE FEATURE CONTRIBUTIONS USING OUT-OF-DIST...

引用

9th International Conference on Learning Representations, ICLR 2021

作者： Lee, Saehyung Park, Changhwa Lee, Hyungyu Yi, Jihun Lee, Jonghyun Yoon, Sungroh Electrical and Computer Engineering AIIS ASRI INMC Institute of Engineering Research Seoul National University Seoul08826 Korea Republic of

Several data augmentation methods deploy unlabeled-in-distribution (UID) data to bridge the gap between the training and inference of neural networks. However, these methods have clear limitations in terms of availability of UID data and dependence of algorithms on pseudo-labels. Herein, we propose a data augmentation method to improve generalization in both adversarial and standard learning by using out-of-distribution (OOD) data that are devoid of the abovementioned issues. We show how to improve generalization theoretically using OOD data in each learning scenario and complement our theoretical analysis with experiments on CIFAR-10, CIFAR-100, and a subset of ImageNet. The results indicate that undesirable features are shared even among image data that seem to have little correlation from a human point of view. We also present the advantages of the proposed method through comparison with other data augmentation methods, which can be used in the absence of UID data. Furthermore, we demonstrate that the proposed method can further improve the existing state-of-the-art adversarial training. © 2021 ICLR 2021 - 9th International Conference on Learning Representations. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

VoiceGuider: Enhancing Out-of-Domain Performance in Parameter-Efficient Speaker-Adaptive Text-to-Speech via Autoguidance

VoiceGuider: Enhancing Out-of-Domain Performance in Paramete...

引用

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

作者： Jiheum Yeom Heeseung Kim Jooyoung Choi Che Hyun Lee Nohil Park Sungroh Yoon Electrical and Computer Engineering Seoul National University Seoul Republic of Korea ECE AIIS ASRI INMC ISRC and IPAI Seoul National University Seoul Republic of Korea

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

When applying parameter-efficient finetuning via LoRA onto speaker adaptive text-to-speech models, adaptation performance may decline compared to full-finetuned counterparts, especially for out-of-domain speakers. Here, we propose VoiceGuider, a parameter-efficient speaker adaptive text-to-speech system reinforced with autoguidance to enhance the speaker adaptation performance, reducing the gap against full-finetuned models. We carefully explore various ways of strengthening autoguidance, ultimately finding the optimal strategy. VoiceGuider as a result shows robust adaptation performance especially on extreme out-of-domain speech data. We provide audible samples in our demo page.

关键词： Degradation Adaptation models Adaptive systems Buildings Signal processing Acoustics Text to speech Speech processing

来源：评论

学校读者我要写书评

暂无评论

Topological Semantic Graph Memory for Image-Goal Navigation

arXiv

引用

arXiv 2022年

作者： Kim, Nuri Kwon, Obin Yoo, Hwiyeon Choi, Yunho Park, Jeongho Oh, Songhwai Department of Electrical and Computer Engineering ASRI Seoul National University Korea Republic of

A novel framework is proposed to incrementally collect landmark-based graph memory and use the collected memory for image goal navigation. Given a target image to search, an embodied robot utilizes semantic memory to find the target in an unknown environment. In this paper, we present a topological semantic graph memory (TSGM), which consists of (1) a graph builder that takes the observed RGB-D image to construct a topological semantic graph, (2) a cross graph mixer module that takes the collected nodes to get contextual information, and (3) a memory decoder that takes the contextual memory as an input to find an action to the target. On the task of an image goal navigation, TSGM significantly outperforms competitive baselines by +5.0-9.0% on the success rate and +7.0-23.5% on SPL, which means that the TSGM finds efficient paths. Additionally, we demonstrate our method on a mobile robot in real-world image goal scenarios. © 2022, CC BY-NC-SA.

关键词： Navigation

来源：评论

学校读者我要写书评

暂无评论

Trust region-based safe distributional reinforcement learning for multiple constraints 23

Trust region-based safe distributional reinforcement learnin...

引用

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Dohyeong Kim Kyungjae Lee Songhwai Oh Dep. of Electrical and Computer Engineering and ASRI Seoul National University Artificial Intelligence Graduate School Chung-Ang University

In safety-critical robotic tasks, potential failures must be reduced, and multiple constraints must be met, such as avoiding collisions, limiting energy consumption, and maintaining balance. Thus, applying safe reinforcement learning (RL) in such robotic tasks requires to handle multiple constraints and use risk-averse constraints rather than risk-neutral constraints. To this end, we propose a trust region-based safe RL algorithm for multiple constraints called a safe distributional actor-critic (SDAC). Our main contributions are as follows: 1) introducing a gradient integration method to manage infeasibility issues in multi-constrained problems, ensuring theoretical convergence, and 2) developing a TD(λ) target distribution to estimate risk-averse constraints with low biases. We evaluate SDAC through extensive experiments involving multi- and single-constrained robotic tasks. While maintaining high scores, SDAC shows 1.93 times fewer steps to satisfy all constraints in multi-constrained tasks and 1.78 times fewer constraint violations in single-constrained tasks compared to safe RL baselines. Code is available at: https://***/rllab-snu/Safe-Distributional-Actor-Critic.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Wasserstein Distributionally Robust Control of Partially Observable Linear Systems: Tractable Approximation and Performance Guarantee

Wasserstein Distributionally Robust Control of Partially Obs...

引用

IEEE Conference on Decision and Control

作者： Astghik Hakobyan Insoon Yang Department of Electrical and Computer Engineering and ASRI Seoul National University Seoul South Korea

ISBN: (数字)9781665467612

ISBN: (纸本)9781665467629

Wasserstein distributionally robust control (WDRC) is an effective method for addressing inaccurate distribution information about disturbances in stochastic systems. It provides various salient features, such as an out-of-sample performance guarantee, while most of the existing methods use full-state observations. In this paper, we develop a computationally tractable WDRC method for discrete-time partially observable linear-quadratic (LQ) control problems. The key idea is to reformulate the WDRC problem as a novel minimax control problem with an approximate Wasserstein penalty. We derive a closed-form expression of the optimal control policy of the approximate problem using a nontrivial Riccati equation. We further show the guaranteed cost property of the resulting controller and identify a provable bound for the optimality gap. Finally, we evaluate the performance of our method through numerical experiments using both Gaussian and non-Gaussian disturbances.

关键词： Robust control Linear systems Costs Closed-form solutions Stochastic systems Riccati equations Optimal control

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：