检索结果-内蒙古大学图书馆

Adaptive test case selection for DNN-based perception functions

学校读者我要写书评

暂无评论

Adaptive test case selection for DNN-based perception functi...

IEEE International Symposium on Systems Engineering (ISSE)

作者： Johannes Bernhard Thomas Schulik Mark Schutera Eric Sax Institute for Information Processing Technologies Karlsruhe Institue of Technology Germany Research and Development ZF Friedrichshafen AG Institute for Automation and Applied Informatics Karlsruhe Institue of Technology Germany

The progress in deep learning methods has bolstered the development of automated vehicles during the last decade. However, the deployment of deep learning methods in safety-critical applications raised questions on their safety. Like other vehicle components, a testing process has to prove the reliability of perception systems. Scalability issues arise when using real-world data to validate perception algorithms due to the immense amount of sensor data that needs to be tested. Simulation tools can complement this testing process, as they can fabricate synthetic data based on variable specifications of test cases. While simulative tools can produce a vast amount of test data, at some point, the testing process is limited by the available computing resources. Identifying test specifications that pose risks to the perception algorithms is crucial for efficiently utilizing these computing resources and estimating functional reliability. We present a pipeline for adaptive test case selection to expose the faults of a deep learning system using synthetic image data generated by a simulation framework. We apply our concept to a state-of-the-art object detector and implement multiple adaptive sampling strategies to demonstrate their ability of early fault detection. Our experiments show that our pipeline can achieve a 95% coverage of system faults while reducing the number of test executions by 90%.

关键词： Deep learning Adaptive systems Computational modeling Scalability Pipelines Tools Data models

Evaluating contextualized language models for Hungarian

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ács, Judit Lévai, Dániel Nemeskey, Dávid Márk Kornai, András Department of Automation and Applied Informatics Budapest University of Technology and Economics Hungary Institute for Computer Science and Control Hungary Alfréd Rényi Institute of Mathematics Hungary

We present an extended comparison of contextualized language models for Hungarian. We compare huBERT, a Hungarian model against 4 multilingual models including the multilingual BERT model. We evaluate these models through three tasks, morphological probing, POS tagging and NER. We find that huBERT works better than the other models, often by a large margin, particularly near the global optimum (typically at the middle layers). We also find that huBERT tends to generate fewer subwords for one word and that using the last subword for token-level tasks is generally a better choice than using the first one. Copyright © 2021, The Authors. All rights reserved.

关键词： Computational linguistics

On the isolated calmness property of implicitly defined multifunctions

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Gfrerer, Helmut Outrata, Jiří V. Institute of Computational Mathematics Johannes Kepler University Linz LinzA-4040 Austria Institute of Information Theory and Automation Czech Academy of Sciences Prague18208 Czech Republic Centre for Informatics and Applied Optimization Federation University of Australia POB 663 BallaratVIC3350 Australia

The paper deals with an extension of the available theory of SCD (subspace containing derivatives) mappings to mappings between spaces of different dimensions. This extension enables us to derive workable sufficient conditions for the isolated calmness of implicitly defined multifunctions around given reference points. This stability property differs substantially from isolated calmness at a point and, possibly in conjunction with the Aubin property, offers a new useful stability concept. The application area includes a broad class of parameterized generalized equations, where the respective conditions ensure a rather strong type of Lipschitztan behavior of their solution *** Codes 65K10, 65K15, 90C33 © 2022, CC BY.

关键词： Mapping

Integrating battery aging in the optimization for bidirectional charging of electric vehicles

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Schwenk, Karl Meisenbacher, Stefan Briegel, Benjamin Harr, Tim Hagenmeyer, Veit Mikut, Ralf Mercedes-Benz AG Sindelfingen Germany Institute for Automation and Applied Informatics Karlsruhe Institute of Technology Karlsruhe Germany Institute for Automation and Applied Informatics Karlsruhe Institute of Technology Karlsruhe Germany

Smart charging of Electric Vehicles (EVs) reduces operating costs, allows more sustainable battery usage, and promotes the rise of electric mobility. In addition, bidirectional charging and improved connectivity enables efficient power grid support. Today, however, uncoordinated charging, e.g. governed by users' habits, is still the norm. Thus, the impact of upcoming smart charging applications is mostly unexplored. We aim to estimate the expenses inherent with smart charging, e.g. battery aging costs, and give suggestions for further research. Using typical on-board sensor data we concisely model and validate an EV battery. We then integrate the battery model into a realistic smart charging use case and compare it with measurements of real EV charging. The results show that i) the temperature dependence of battery aging requires precise thermal models for charging power greater than 7 kW, ii) disregarding battery aging underestimates EVs' operating costs by approx. 30%, and iii) the profitability of Vehicle-to-Grid (V2G) services based on bidirectional power flow, e.g. energy arbitrage, depends on battery aging costs and the electricity price spread. © 2020, CC BY.

关键词： Neural networks

Field Experiments with an Automated Utility Platform for Transportation and Work Processes

学校读者我要写书评

暂无评论

Field Experiments with an Automated Utility Platform for Tra...

IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC)

作者： Raimund Edlinger Stefan Dumberger Roman Froschauer Wolfgang Pointner Marcel Zeilinger Andreas Nü chter Smart Automation and Robotics University of Applied Sciences Upper Austria Wels Austria Assistive &#x0026 Autonomous Systems AIT &#x2013 Austrian Institute of Technology Vienna Austria Informatics VII - Robotics and Telematics Julius-Maximilians-University W&#x00FC rzburg W&#x00FC rzburg Germany

The automation of transport and work tasks in various application domains continues to be a major challenge. The proposed paper presents key technologies within a process from sensing to reasoning and includes: (i) understanding the scene and its relations through new approaches in object classification and pose estimation based on machine learning, as well as methods for reliable detection of objects and object features; (ii) reliable and safe outdoor localization and mapping based on multi-modal approaches and innovative camera-based methods. The fusion of different sensor data is used for reliable obstacle detection and collision avoidance. The investigation of suitable system architectures, data processing platforms and communication mechanisms taking into account safety requirements is a prerequisite for the operation in a safety-critical environment. The paper provides a proof of concept for an automated platform to demonstrate and evaluate application scenario provided by the Linz Airport in the form of an automated transport process for air freight between a terminal and the aircraft.

关键词： Navigation Transportation Systems architecture Switches Robot sensing systems Software Software reliability

Computer modeling application of fluid outflow from vessels

学校读者我要写书评

暂无评论

Materials Science Forum 2019年 952卷 250-257页

作者： Kraváriková, Helena Faculty of Materials Science and Technology in Trnava Institute of Applied Informatics Automation and Mechatronics Bottova 26 Trnava917 24 Slovakia

The aim of the paper is to evaluate numerical analysis of the fluid flow during the outflow from vessel orifices at various locations. The problems of the outflow velocities and pressure fields were well-chosen for the given purposes. The selected fluid flow problems were solved by numerical simulation using FEM in ANSYS. For numerical simulation, we used the basic steps to design an abstract model in the ANSYS virtual environment. Numerical simulation requires a geometric model complemented by physical properties of flowing fluids as well as both the initial and boundary conditions. It is then possible to calculate the velocity and pressure fields by numerical simulation for a particular fluid type. The results obtained from the numerical simulation were compared with those of the analytical solution. The results obtained from modeling and numerical simulation correspond to the actual values with minimum deviations. The demonstrated type of the problem solved by numerical simulation and modeling confirmed the advantages and possibilities of flexible solutions for any combination of problems in the field of fluid dynamics. Modeling and numerical simulation of fluid flow can provide results regarding the speed and the pressure fields in vessels and pipelines. © 2019 Trans Tech Publications Ltd, Switzerland.

关键词： Orifices

Automated File Labeling for Heterogeneous Files Organization Using Machine Learning

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2023年第2期74卷 3263-3278页

作者： Sagheer Abbas Syed Ali Raza MAKhan Muhammad Adnan Khan Atta-ur-Rahman Kiran Sultan Amir Mosavi School of Computer Science National College of Business Administration&EconomicsLahore54000Pakistan Department of Computer Science GC University LahorePakistan Riphah School of Computing&Innovation Faculty of ComputingRiphah International UniversityLahore CampusLahore54000Pakistan Department of Software Pattern Recognition and Machine Learning LabGachon UniversitySeongnam13120Korea Department of Computer Science College of Computer Science and Information Technology(CCSIT)Imam Abdulrahman Bin Faisal University(IAU)P.O.Box 1982Dammam31441Saudi Arabia Department of CIT The Applied CollegeKing Abdulaziz UniversityJeddah31261Saudi Arabia John von Neumann Faculty of Informatics Obuda UniversityBudapest1034Hungary Institute of Information Engineering Automation and MathematicsSlovak University of Technology in BratislavaBratislava81107Slovakia Faculty of Civil Engineering TU-DresdenDresden01062Germany

File labeling techniques have a long history in analyzing the anthological trends in computational *** situation becomes worse in the case of files downloaded into systems from the ***,most users either have to change file names manually or leave a meaningless name of the files,which increases the time to search required files and results in redundancy and duplications of user ***,no significant work is done on automated file labeling during the organization of heterogeneous user files.A few attempts have been made in topic ***,one major drawback of current topic modeling approaches is better *** rely on specific language types and domain similarity of the *** this research,machine learning approaches have been employed to analyze and extract the information from heterogeneous corpus.A different file labeling technique has also been used to get the meaningful and`cohesive topic of the *** results show that the proposed methodology can generate relevant and context-sensitive names for heterogeneous data files and provide additional insight into automated file labeling in operating systems.

关键词： Automated file labeling file organization machine learning topic modeling

Long-term On-line Identification of Time-varying Systems

学校读者我要写书评

暂无评论

Long-term On-line Identification of Time-varying Systems

International Carpathian Control Conference (ICCC)

作者： Ján Vachálek Dana Šišmišová Ivan Fiťka Matej Šimovec Faculty of Mechanical Engineering Institute of Automation Measurement and Applied Informatics Slovak University of Technology in Bratislava Nám. Slobody 17 Bratislava Slovak Republic

This article discusses the possibility of deploying regression recursive continuous identification methods, which are using regularized exponential forgetting, during long-term monitoring of time-varying systems. The emphasis is placed on the long-term deployment since the advantage of this algorithm will only become evident with long-term deployment. These continuous identification algorithms are going to be applied on the mathematical model of the identified system for the needs of adaptive control. The article presents the implemented continuous identification techniques, where the non-informative data, that can possibly destabilize the numerical calculations of the identified system parameters, are weighted by the Dyadic reduction algorithm. These techniques are using an alternative covariance matrix for its “forgetting” in order to maintain the initial system dynamics in the mathematical model of the identified system. The matrix also suppresses the impact of high amounts of non-informative data received from the long-term operation the industrial systems with slow dynamics. The focus is mainly set on the non-informative data, which emerge especially during the long-term run of the tuned systems operating in between the bounds of the selected working point. A “Motor-Flywheel” laboratory model is used for the validation tests of the modified and the standard regression recursive algorithms.

关键词： System dynamics Heuristic algorithms Prediction algorithms Numerical models Mathematical model Reliability Covariance matrices

The strong effect of network resolution on electricity system models with high shares of wind and solar

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Frysztacki, Martha Maria Hörsch, Jonas Hagenmeyer, Veit Brown, Tom Institute for Automation and Applied Informatics Karlsruhe Institute of Technology Eggenstein-Leopoldshafen76344 Germany Frankfurt Institute for Advanced Studies Ruth-Moufang-Straße 1 Frankfurt am Main60438 Germany

Energy system modellers typically choose a low spatial resolution for their models based on administrative boundaries such as countries, which eases data collection and reduces computation times. However, a low spatial resolution can lead to sub-optimal investment decisions for wind and solar generation. Ignoring power grid bottlenecks within regions tends to underestimate system costs, while combining locations with different wind and solar capacity factors in the same resource class tends to overestimate costs. We investigate these two competing effects in a capacity expansion model for Europe’s power system with a high share of renewables, taking advantage of newly-available high-resolution datasets as well as computational advances. We vary the number of nodes, interpolating between a 37-node model based on country and synchronous zone boundaries, and a 1024-node model based on the location of electricity substations. If we focus on the effect of renewable resource resolution and ignore network restrictions, we find that a higher resolution allows the optimal solution to concentrate wind and solar capacity at sites with better capacity factors and thus reduces system costs by up to 10% compared to a low resolution model. This results in a big swing from offshore to onshore wind investment. However, if we introduce grid bottlenecks by raising the network resolution, costs increase by up to 23% as generation has to be sourced more locally at sites with worse capacity factors. These effects are most pronounced in scenarios where grid expansion is limited, for example, by low local acceptance. We show that allowing grid expansion mitigates some of the effects of the low grid resolution, and lowers overall costs by around 16%. © 2021, CC BY.

关键词： Investments