检索结果-内蒙古大学图书馆

[DC] Towards Comprehensible and Expressive Teleportation Techniques in Immersive Virtual Environments

学校读者我要写书评

暂无评论

[DC] Towards Comprehensible and Expressive Teleportation Tec...

Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW), IEEE Conference on

作者： Daniel Rupp Visual Computing Institute RWTH Aachen University

ISBN: (数字)9798331514846

ISBN: (纸本)9798331525637

Teleportation, a popular navigation technique in virtual environments, is favored for its efficiency and reduction of cybersickness but presents challenges such as reduced spatial awareness and limited navigational freedom compared to continuous techniques. I would like to focus on three aspects that advance our understanding of teleportation in both the spatial and the temporal domain. 1) An assessment of different parametrizations of common mathematical models used to specify the target location of the teleportation and the influence on teleportation distance and accuracy. 2) Extending teleportation capabilities to improve navigational freedom, comprehensibility, and accuracy. 3) Adapt teleportation to the time domain, mediating temporal disorientation. The results will enhance the expressivity of existing teleportation interfaces and provide validated alternatives to their steering-based counterparts.

关键词： Accuracy Three-dimensional displays Navigation Cybersickness Conferences Virtual environments Teleportation User interfaces Mathematical models Time-domain analysis

Using (Not so) Large Language Models for Generating Simulation Models in a Formal DSL - A Study on Reaction Networks

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Kreikemeyer, Justin N. Jankowski, Milosz Wilsdorf, Pia Uhrmacher, Adelinde M. Institute for Visual and Analytic Computing University of Rostock Rostock Germany

Formal languages are an integral part of modeling and simulation. They allow the distillation of knowledge into concise simulation models amenable to automatic execution, interpretation, and analysis. However, the arguably most humanly accessible means of expressing models is through natural language, which is not easily interpretable by computers. Here, we evaluate how a Large Language Model (LLM) might be used for formalizing natural language into simulation models. Existing studies only explored using very large LLMs, like the commercial GPT models, without fine-tuning model weights. To close this gap, we show how an open-weights, 7B-parameter Mistral model can be fine-tuned to translate natural language descriptions to reaction network models in a domain-specific language, offering a self-hostable, compute-, and memory efficient alternative. To this end, we develop a synthetic data generator to serve as the basis for fine-tuning and evaluation. Our quantitative evaluation shows that our fine-tuned Mistral model can recover the ground truth simulation model in up to 84.5% of cases. In addition, our small-scale user study demonstrates the model’s practical potential for one-time generation as well as interactive modeling in various domains. While promising, in its current form, the fine-tuned small LLM cannot catch up with large LLMs. We conclude that higher-quality training data are required, and expect future small and open-source LLMs to offer new opportunities. © 2025, CC BY.

关键词： Modeling languages

Diffusion at Absolute Zero: Langevin Sampling Using Successive Moreau Envelopes

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Habring, Andreas Falk, Alexander Pock, Thomas Institute of Visual Computing Graz University of Technology Graz Austria

In this article we propose a novel method for sampling from Gibbs distributions of the form π(x) ∝ exp(−U(x)) with a potential U(x). In particular, inspired by diffusion models we propose to consider a sequence (πtk)k of approximations of the target density, for which πtk ≈ π for k small and, on the other hand, πtk exhibits favorable properties for sampling for k large. This sequence is obtained by replacing parts of the potential U by its Moreau envelopes. Sampling is performed in an Annealed Langevin type procedure, that is, sequentially sampling from πtk for decreasing k, effectively guiding the samples from a simple starting density to the more complex target. In addition to a theoretical analysis we show experimental results supporting the efficacy of the method in terms of increased convergence speed and applicability to multi-modal densities π.MSC Codes 65C40, 65C05, 68U10, 65C60 © 2025, CC BY.

关键词： Markov chains

Learning Fine-to-Coarse Cuboid Shape Abstraction

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Kobsik, Gregor Henkel, Morten He, Yanjiang Czech, Victor Elsner, Tim Lim, Isaak Kobbelt, Leif Visual Computing Institute RWTH Aachen University Aachen Germany

The abstraction of 3D objects with simple geometric primitives like cuboids allows to infer structural information from complex geometry. It is important for 3D shape understanding, structural analysis and geometric modeling. We introduce a novel fine-to-coarse unsupervised learning approach to abstract collections of 3D shapes. Our architectural design allows us to reduce the number of primitives from hundreds (fine reconstruction) to only a few (coarse abstraction) during training. This allows our network to optimize the reconstruction error and adhere to a user-specified number of primitives per shape while simultaneously learning a consistent structure across the whole collection of data. We achieve this through our abstraction loss formulation which increasingly penalizes redundant primitives. Furthermore, we introduce a reconstruction loss formulation to account not only for surface approximation but also volume preservation. Combining both contributions allows us to represent 3D shapes more precisely with fewer cuboid primitives than previous work. We evaluate our method on collections of man-made and humanoid shapes comparing with previous state-of-the-art learning methods on commonly used benchmarks. Our results confirm an improvement over previous cuboid-based shape abstraction techniques. Furthermore, we demonstrate our cuboid abstraction in downstream tasks like clustering, retrieval, and partial symmetry detection. Copyright © 2025, The Authors. All rights reserved.

关键词： Unsupervised learning

[DC] Conversational Agents for Natural Human-Computer Interaction in Virtual Reality

学校读者我要写书评

暂无评论

[DC] Conversational Agents for Natural Human-Computer Intera...

Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW), IEEE Conference on

作者： Matteo Bosco Institute of Visual Computing and Human-Centered Technology TU Wien Austria

ISBN: (数字)9798331514846

ISBN: (纸本)9798331525637

With the emergence of large language models (LLMs), conversational agents (CAs) have gained significant attention across various domains, including virtual reality (VR). Although LLMs are capable of simulating human dialogue with impressive realism, interactions with LLM-powered CAs still appear unnatural and lack a human-like feel. This is primarily due to several key challenges, including limitations in the accuracy of speech recognition and the quality of speech synthesis. Furthermore, CAs need more autonomy and more natural turn-taking abilities to enable realistic and engaging interactions in VR. This position paper outlines my PhD research by discussing my motivation for improving CAs in VR environments, summarizing my previous work, and presenting my planned research directions.

关键词： Human computer interaction Three-dimensional displays Large language models Conferences Natural languages Virtual reality Speech recognition Medical services Companies Speech synthesis

GenAI as a Tool for Content Generation in Hypermedia Edutainment Applications: Potential and Limitations 17

学校读者我要写书评

暂无评论

GenAI as a Tool for Content Generation in Hypermedia Edutain...

17th International Conference on Computer Supported Education, CSEDU 2025

作者： Heiden, Wolfgang Saitova, Veronika Kless, Tea Wegner, Valeska Rötter, David Neteler, Thomas Institute of Visual Computing Bonn-Rhein-Sieg University of Applied Sciences Sankt Augustin Germany Department of Mathematics Justus-Liebig-Universität Gießen Gießen Germany

ISBN: (纸本)9789897587467

Generative AI can considerably speed up the process of producing narrative content including different media. This may be particularly helpful for the generation of modular variations on narrative themes in hypermedia, crossmedia, or transmedia contexts, thereby enabling personalized access to the content by heterogenous target groups. We present an example where GenAI has been applied for image creation and translation of a text to multiple languages for a crossmedia edutainment project transferring IT security knowledge to vulnerable groups. GenAI still seems inadequate to produce interesting narrative text integrating dedicated educational content. AI-generated illustrations often require manual rework. However, LLM support in multilingual translations displays more intelligent solutions than expected, including the implementation of a password generation process from a narrated description. Copyright © 2025 by SCITEPRESS - Science and Technology Publications, Lda.

关键词： Hypermedia systems

EVLoc: Event-based visual Localization in LiDAR Maps via Event-Depth Registration

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Chen, Kuangyi Zhang, Jun Fraundorfer, Friedrich Institute of Visual Computing Graz University of Technology Graz8010 Austria

Event cameras are bio-inspired sensors with some notable features, including high dynamic range and low latency, which makes them exceptionally suitable for perception in challenging scenarios such as high-speed motion and extreme lighting conditions. In this paper, we explore their potential for localization within pre-existing LiDAR maps, a critical task for applications that require precise navigation and mobile manipulation. Our framework follows a paradigm based on the refinement of an initial pose. Specifically, we first project LiDAR points into 2D space based on a rough initial pose to obtain depth maps, and then employ an optical flow estimation network to align events with LiDAR points in 2D space, followed by camera pose estimation using a PnP solver. To enhance geometric consistency between these two inherently different modalities, we develop a novel frame-based event representation that improves structural clarity. Additionally, given the varying degrees of bias observed in the ground truth poses, we design a module that predicts an auxiliary variable as a regularization term to mitigate the impact of this bias on network convergence. Experimental results on several public datasets demonstrate the effectiveness of our proposed method. To facilitate future research, both the code and the pre-trained models are made available online1 Copyright © 2025, The Authors. All rights reserved.

关键词： Optical flows

Semantic-oriented visual Prompt Learning for Class Incremental Learning

学校读者我要写书评

暂无评论

Semantic-oriented Visual Prompt Learning for Class Increment...

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

作者： Guo, Shuai Gu, Yang Ma, Yuan Zhang, Yingwei Weng, Weining Liu, Jun Dai, Weiwei Chen, Yiqiang Beijing Key Laboratory of Mobile Computing and Pervasive Device Institute of Computing Technology Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Institute of Digital Ophthalmology and Visual Science Changsha Aier Eye Hospital Changsha China Peng Cheng Laboratory Shenzhen China

ISBN: (纸本)9798350368741

Class-incremental learning (CIL) enables models to continuously learn new classes while addressing catastrophic forgetting. With the introduction of pre-trained models, new tuning paradigms have emerged for CIL. This paper revisits parameter-efficient fine-tuning (PEFT) methods in the context of incremental learning. Prior studies reveal that PEFT methods' extended parameters do not directly contribute to semantic perception, limiting performance with significant category and domain gaps. To address this, we propose semantic-oriented visual prompt learning (SVPL), which enhances semantic perception and improves task-specific knowledge extraction. SVPL assigns learnable prompts to each class, using a contrastive group alignment to align prompts to task-specific semantic spaces, thus preserving relationships between old and new knowledge. Additionally, hierarchical semantic delivery allows the semantic transformation of prompt groups from shallow to deep layers to facilitate efficient knowledge mining and enable effective learning of new knowledge. Extensive experimental results on five benchmarks demonstrate the superior performance of our methods. © 2025 IEEE.

关键词： catastrophic forgetting class incremental learning pre-trained model prompt learning vision transformer

An Investigation of Beam Density on LiDAR Object Detection Performance

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Griesbacher, Christoph Fruhwirth-Reisinger, Christian Institute of Visual Computing TU Graz Austria Christian Doppler Laboratory for Embedded Machine Learning Austria

Accurate 3D object detection is a critical component of autonomous driving, enabling vehicles to perceive their surroundings with precision and make informed decisions. LiDAR sensors, widely used for their ability to provide detailed 3D measurements, are key to achieving this capability. However, variations between training and inference data can cause significant performance drops when object detection models are employed in different sensor settings. One critical factor is beam density, as inference on sparse, cost-effective LiDAR sensors is often preferred in real-world applications. Despite previous work addressing the beam-density-induced domain gap, substantial knowledge gaps remain, particularly concerning dense 128-beam sensors in cross-domain scenarios. To gain better understanding of the impact of beam density on domain gaps, we conduct a comprehensive investigation that includes an evaluation of different object detection architectures. Our architecture evaluation reveals that combining voxel- and point-based approaches yields superior cross-domain performance by leveraging the strengths of both representations. Building on these findings, we analyze beam-density-induced domain gaps and argue that these domain gaps must be evaluated in conjunction with other domain shifts. Contrary to conventional beliefs, our experiments reveal that detectors benefit from training on denser data and exhibit robustness to beam density variations during inference. © 2025, CC BY-NC-SA.

关键词： Object detection