检索结果-内蒙古大学图书馆

Low Fidelity Digital Twin for Automated Driving Systems: Use Cases and Automatic Generation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Vlasak, Jiri Klapálek, Jaroslav Kollarčík, Adam Sojka, Michal Hanzálek, Zdeněk Faculty of Electrical Engineering Czech Technical University in Prague Czech Republic Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague Czech Republic

Automated driving systems are an integral part of the automotive industry. Tools such as Robot Operating System and simulators support their development. However, in the end, the developers must test their algorithms on a real vehicle. To better observe the difference between reality and simulation–the reality gap–digital twin technology offers real-time communication between the real vehicle and its model. We present a low fidelity digital twin generator and describe situations where automatic generation is preferable to high fidelity simulation. We validate our approach of generating a virtual environment with a vehicle model by replaying the data recorded from the real vehicle. © 2024, CC BY-NC-ND.

关键词： Automotive industry

Thinking Tokens for Language Modeling

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Herel, David Mikolov, Tomas Faculty of Electrical Engineering Czech Technical University in Prague Czech Republic Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague Czech Republic

How much is 56 times 37? Language models often make mistakes in these types of difficult calculations. This is usually explained by their inability to perform complex reasoning. Since language models rely on large training sets and great memorization capability, naturally they are not equipped to run complex calculations. However, one can argue that humans also cannot perform this calculation immediately and require a considerable amount of time to construct the solution. In order to enhance the generalization capability of language models, and as a parallel to human behavior, we propose to use special’thinking tokens’ which allow the model to perform much more calculations whenever a complex problem is encountered. © 2024, CC BY.

关键词： Modeling languages

WildGaussians: 3D Gaussian splatting in the wild 24

学校读者我要写书评

暂无评论

WildGaussians: 3D Gaussian splatting in the wild

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Jonas Kulhanek Songyou Peng Zuzana Kukelova Marc Pollefeys Torsten Sattler Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague and Faculty of Electrical Engineering Czech Technical University in Prague and Department of Computer Science ETH Zurich Department of Computer Science ETH Zurich Visual Recognition Group Faculty of Electrical Engineering Czech Technical University in Prague Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague

ISBN: (纸本)9798331314385

While the field of 3D scene reconstruction is dominated by NeRFs due to their photorealistic quality, 3D Gaussian Splatting (3DGS) has recently emerged, offering similar quality with real-time rendering speeds. However, both methods primarily excel with well-controlled 3D scenes, while in-the-wild data - characterized by occlusions, dynamic objects, and varying illumination - remains challenging. NeRFs can adapt to such conditions easily through per-image embedding vectors, but 3DGS struggles due to its explicit representation and lack of shared parameters. To address this, we introduce WildGaussians, a novel approach to handle occlusions and appearance changes with 3DGS. By leveraging robust DINO features and integrating an appearance modeling module within 3DGS, our method achieves state-of-the-art results. We demonstrate that WildGaussians matches the real-time rendering speed of 3DGS while surpassing both 3DGS and NeRF baselines in handling in-the-wild data, all within a simple architectural framework.

关键词：

D-InLoc++: Indoor Localization in Dynamic Environments 44th

学校读者我要写书评

暂无评论

D-InLoc++: Indoor Localization in Dynamic Environments

44th DAGM German Conference on Pattern Recognition, DAGM GCPR 2022

作者： Dubenova, Martina Zderadickova, Anna Kafka, Ondrej Pajdla, Tomas Polic, Michal CIIRC - Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague Prague Czech Republic Visual Recognition Group Faculty of Electrical Engineering Czech Technical University in Prague Prague Czech Republic

ISBN: (纸本)9783031167874

Most state-of-the-art localization algorithms rely on robust relative pose estimation and geometry verification to obtain moving object agnostic camera poses in complex indoor environments. However, this approach is prone to mistakes if a scene contains repetitive structures, e.g., desks, tables, boxes, or moving people. We show that the movable objects incorporate non-negligible localization error and present a new straightforward method to predict the six-degree-of-freedom (6DoF) pose more robustly. We equipped the localization pipeline InLoc with real-time instance segmentation network YOLACT++. The masks of dynamic objects are employed in the relative pose estimation step and in the final sorting of camera pose proposal. At first, we filter out the matches laying on masks of the dynamic objects. Second, we skip the comparison of query and synthetic images on the area related to the moving object. This procedure leads to a more robust localization. Lastly, we describe and improve the mistakes caused by gradient-based comparison between synthetic and query images and publish a new pipeline for simulation of environments with movable objects from the Matterport scans. All the codes are available on https://***/dubenma/D-InLocpp. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Pipelines

Beyond the Safeguards: Exploring the Security Risks of ChatGPT

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Derner, Erik Batistič, Kristina Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague Prague Czech Republic Ljubljana Slovenia

The increasing popularity of large language models (LLMs) such as ChatGPT has led to growing concerns about their safety, security risks, and ethical implications. This paper aims to provide an overview of the different types of security risks associated with ChatGPT, including malicious text and code generation, private data disclosure, fraudulent services, information gathering, and producing unethical content. We present an empirical study examining the effectiveness of Chat-GPT's content filters and explore potential ways to bypass these safeguards, demonstrating the ethical implications and security risks that persist in LLMs even when protections are in place. Based on a qualitative analysis of the security implications, we discuss potential strategies to mitigate these risks and inform researchers, policymakers, and industry professionals about the complex security challenges posed by LLMs like ChatGPT. This study contributes to the ongoing discussion on the ethical and security implications of LLMs, underscoring the need for continued research in this area. © 2023, CC BY.

关键词： Natural language processing systems

CLFT: Camera-LiDAR Fusion Transformer for Semantic Segmentation in Autonomous Driving

学校读者我要写书评

暂无评论

IEEE Transactions on Intelligent Vehicles

IEEE Transactions on Intelligent Vehicles 2024年

作者： Gu, Junyi Bellone, Mauro Pivoňka, Tomáš Sell, Raivo Tallinn University of Technology Department of Mechanical and Industrial Engineering Estonia Tallinn University of Technology FinEst Centre for Smart Cities Estonia Czech Technical University in Prague Czech Institute of Informatics Robotics and Cybernetics Department of Cybernetics Czech Republic

Critical research about camera-and-LiDAR-based semantic object segmentation for autonomous driving significantly benefited from the recent development of deep learning. Specifically, the vision transformer is the novel ground-breaker that successfully brought the multi-head-attention mechanism to computer vision applications. Therefore, we propose a vision-transformer-based network to carry out camera-LiDAR fusion for semantic segmentation applied to autonomous driving. Our proposal uses the novel progressive-assemble strategy of vision transformers on a double-direction network and then integrates the results in a cross-fusion strategy over the transformer decoder layers. Unlike other works in the literature, our camera-LiDAR fusion transformers have been evaluated in challenging conditions like rain and low illumination, showing robust performance. The paper reports the segmentation results over the vehicle and human classes in different modalities: camera-only, LiDAR-only, and camera-LiDAR fusion. We perform coherent controlled benchmark experiments of the camera-LiDAR fusion transformer (CLFT) against other networks that are also designed for semantic segmentation. The experiments aim to evaluate the performance of CLFT independently from two perspectives: multimodal sensor fusion and backbone architectures. The quantitative assessments show our CLFT networks yield an improvement of up to 10% for challenging dark-wet conditions when comparing with Fully-Convolutional-Neural-Network-based (FCN) camera-LiDAR fusion neural network. Contrasting to the network with transformer backbone but using single modality input, the all-around improvement is 5-10%. Our full code is available online for an interactive demonstration and application. © 2016 IEEE.

关键词： Semantic Segmentation

Electrification of Transportation: A Hybrid Benders/SDDP Algorithm for Optimal Charging Station Trading

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sohrabi, Farnaz Rohaninejad, Mohammad Bemš, Július Hanzálek, Zdeněk Faculty of Electrical Engineering Czech Technical University in Prague Prague Czech Republic Industrial Informatics Department Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague Prague Czech Republic

This paper examines the electrification of transportation as a response to environmental challenges caused by fossil fuels, exploring the potential of battery electric vehicles and hydrogen fuel cell vehicles as alternative solutions. However, a significant barrier to their widespread adoption is the limited availability of charging infrastructure. Therefore, this study proposes the development of comprehensive charging stations capable of accommodating both battery and hydrogen vehicles to address this challenge. The energy is purchased from the day-ahead and intraday auction-based electricity markets, where the electricity price is subject to uncertainty. Therefore, a two-stage stochastic programming model is formulated while the price scenarios are generated utilizing a k-means clustering algorithm. Given the complexity of the proposed model, an efficient solution approach is developed through the hybridization of the Benders decomposition algorithm and stochastic dual dynamic programming. In the Benders master problem, day-ahead bidding variables are determined, whereas the Benders sub-problem addresses intraday bidding and charging station scheduling variables, employing stochastic dual dynamic programming to tackle its intractability. Additionally, we transform the mixed integer linear program model of the second stage problem into a linear program, confirming its validity through KKT conditions. Our model provides practical insights for making informed decisions in electricity markets based on sequential auctions. While the bidding curves submitted to the day-ahead market remain unaffected by scenarios, those submitted to the intra-day market show dependence on fluctuations in day-ahead market prices. © 2024, CC BY-NC-ND.

关键词： Stochastic systems

Optimal Trading of a Charging-Station Company in Auction Markets for Electricity

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Sohrabi, Farnaz Rohaninejad, Mohammad Hesamzadeh, Mohammad Reza Bemš, Július Czech Technical University Prague Czech Republic Department of Industrial Informatics The Institute of Informatics Robotics and Cybernetics Czech Technical University Prague Czech Republic KTH Royal Institute of Technology Sweden Department of Economics Management and Humanitie

This paper addresses a charging-station company (Chargco) for electric and hydrogen vehicles. The optimal trading of the Chargco in day-ahead and intraday auction markets for electricity is modeled as a stochastic Mixed-Integer Quadratic Program (MIQP). We propose a series of linearization and reformulation techniques to reformulate the stochastic MIQP as a mixed-integer linear program (MILP). To model stochasticity, we utilize generative adversarial networks to cluster electricity market price scenarios. Additionally, a combination of random forests and linear regression is employed to model the relationship between Chargco electricity and hydrogen loads and their selling prices. Finally, we propose an Improved L-Shaped Decomposition (ILSD) algorithm to solve our stochastic MILP. Our ILSD algorithm not only addresses infeasibilities through an innovative approach but also incorporates warm starts, valid inequalities and multiple generation cuts, thereby reducing computational complexity. Numerical experiments illustrate the Chargco trading using our proposed stochastic MILP and its solution algorithm. © 2025, CC BY-NC-ND.

关键词： Stochastic programming

Large-scale Pre-training for Grounded Video Caption Generation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Kazakos, Evangelos Schmid, Cordelia Sivic, Josef Czech Institute of Informatics Robotics and Cybernetics The Czech Technical University in Prague Czech Republic Inria École normale supérieure CNRS PSL Research University France

We propose a novel approach for captioning and object grounding in video, where the objects in the caption are grounded in the video via temporally dense bounding boxes. We introduce the following contributions. First, we present a large-scale automatic annotation method that aggregates captions grounded with bounding boxes across individual frames into temporally dense and consistent bounding box annotations. We apply this approach on the HowTo100M dataset to construct a large-scale pretraining dataset, named HowToGround1M. We also introduce a Grounded Video Caption Generation model, dubbed GROVE, and pre-train the model on HowToGround1M. Second, we introduce a new dataset, called iGround, of 3500 videos with manually annotated captions and dense spatio-temporally grounded bounding boxes. This allows us to measure progress on this challenging problem, as well as to fine-tune our model on this small-scale but high-quality data. Third, we demonstrate that our approach achieves state-ofthe-art results on the proposed iGround dataset compared to a number of baselines, as well as on the VidSTG and ActivityNet-Entities datasets. We perform extensive ablations that demonstrate the importance of pre-training using our automatically annotated HowToGround1M dataset followed by fine-tuning on the manually annotated iGround dataset and validate the key technical contributions of our model. © 2025, CC BY-NC-SA.

关键词： Large datasets