检索结果-内蒙古大学图书馆

AI vs. AI: The Detection Game

学校读者我要写书评

暂无评论

AI vs. AI: The Detection Game

IEEE International Forum on Research and Technologies for Society and Industry Leveraging a better tomorrow (RTSI)

作者： Mauro Coccoli Gabriella Patanè DIBRIS - Department of Informatics Bioengineering Robotics and Systems Science University of Genoa Genoa Freelancer Genoa

ISBN: (数字)9798350362138

ISBN: (纸本)9798350362145

In this paper, we present the results achieved by performing a test addressed to evaluate the capabilities of the so-called AI content Detectors or AI Detection systems. Such systems promise to find out whether a given text was written by a human or created by an artificial intelligence (AI). The ability to identify the “synthetic” nature of a text is paramount in many workplaces as well as in the educational field, where we focused our research. More specifically, we considered the academic environment and the frequent request by faculty to students to write term papers on a wide variety of topics. Then, it is essential for teachers to be able to evaluate content quality as well as their authenticity. Until now, anti-plagiarism tools have been widely used to verify that homework assignments were not copied, but the recent advent of generative artificial intelligence-based tools changed the game, moving the focus to automatic generation. In both cases, we are facing a violation of the educational pact by a student, versus a teacher, if not outright an offence. Conversely, an unfounded accusation of fraud by a teacher toward a student would be equally serious. Therefore, to use automatic detection tools one must have a guarantee on the reliability of their output.

关键词： Training Analytical models Technological innovation Generative AI Games Writing Transformers Software reliability Artificial intelligence Sustainable development

Robustness-Congruent Adversarial Training for Secure Machine Learning Model Updates

学校读者我要写书评

暂无评论

IEEE Transactions on Pattern Analysis and Machine Intelligence 2025年 PP卷 PP页

作者： Angioni, Daniele Demetrio, Luca Pintor, Maura Oneto, Luca Anguita, Davide Biggio, Battista Roli, Fabio University of Cagliari Department of Electrical and Electronic Engineering Italy University of Genova Department of Informatics Bioengineering Robotics and Systems Engineering Italy

Machine-learning models demand periodic updates to improve their average accuracy, exploiting novel architectures and additional data. However, a newly updated model may commit mistakes that the previous model did not make. Such misclassifications are referred to as negative flips, experienced by users as a regression of performance. In this work, we show that this problem also affects robustness to adversarial examples, hindering the development of secure model update practices. In particular, when updating a model to improve its adversarial robustness, previously ineffective adversarial attacks on some inputs may become successful, causing a regression in the perceived security of the system. We propose a novel technique, named robustness-congruent adversarial training, to address this issue. It amounts to fine-tuning a model with adversarial training, while constraining it to retain higher robustness on the samples for which no adversarial example was found before update. We show that our algorithm and, more generally, learning with non-regression constraints, provides a theoretically-grounded framework to train consistent estimators. Our experiments on robust models for computer vision confirm that both accuracy and robustness, even if improved after model update, can be affected by negative flips, and our robustness-congruent adversarial training can mitigate the problem, outperforming competing baseline methods. © 1979-2012 IEEE.

关键词： Adversarial Examples Adversarial Robustness Machine Learning Regression Testing

RV4Chatbot: Are Chatbots Allowed to Dream of Electric Sheep?

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Gatti, Andrea Mascardi, Viviana Ferrando, Angelo Department of Informatics Bioengineering Robotics and Systems Engineering University of Genoa Genoa Italy Department of Physics Informatics and Mathematics University of Modena and Reggio Emilia Modena Italy

Chatbots have become integral to various application domains, including those with safety-critical considerations. As a result, there is a pressing need for methods that ensure chatbots consistently adhere to expected, safe behaviours. In this paper, we introduce RV4Chatbot, a Runtime Verification framework designed to monitor deviations in chatbot behaviour. We formalise expected behaviours as interaction protocols between the user and the chatbot. We present the RV4Chatbot design and describe two implementations that instantiate it: RV4Rasa, for monitoring chatbots created with the Rasa framework, and RV4Dialogflow, for monitoring Dialogflow chatbots. Additionally, we detail experiments conducted in a factory automation scenario using both RV4Rasa and RV4Dialogflow. © 2024, CC BY.

关键词： Bot (Internet)

Design, Implementation and Testing of an EtherCAT-based Network for Multi-modal Distributed Sensing Architectures

学校读者我要写书评

暂无评论

Design, Implementation and Testing of an EtherCAT-based Netw...

IEEE International Conference on Automation Science and engineering (CASE)

作者： Francesco Giovinazzo Alessandro Perri Marco Staiano Francesco Grella Marco Sartore Manuela Adami Riccardo Galletti Giorgio Cannata Department of Informatics Bioengineering Robotics and Systems Engineering (DIBRIS) Università di Genova Genova Italy ElbaTech Srl Marciana Italy

ISBN: (数字)9798350358513

ISBN: (纸本)9798350358520

This paper introduces an EtherCAT-based communication infrastructure designed for large-area multi-modal sensor networks in robotic applications. The system addresses critical application-oriented challenges, such as bandwidth optimization, cable routing efficiency, network scalability and fault tolerance. The proposed architecture is experimentally validated with ProxySKIN, a distributed sensing technology integrating arrays of tactile and proximity sensors, designed to cover large areas of robotic manipulators. The implemented solution leverages a redundant EtherCAT network driven by custom-made miniaturized slave devices to collect comprehensive data regarding the robot’s surrounding space. Bench-test experiments conducted on over 6400 pressure sensors and 20 multi-zone proximity sensors show reliable data collection with minimal communication latency through a unique EtherCAT backbone.

关键词： Pressure sensors Scalability Robot sensing systems Manipulators Routing Real-time systems Sensors Optimization Sensor arrays Testing

A Systematic Review on Custom Data Gloves

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Belcamino, Valerio Carfì, Alessandro Mastrogiovanni, Fulvio TheEngineRoom Department of Informatics Bioengineering Robotics and Systems Engineering University of Genoa Via Opera Pia 13 Genoa16145 Italy

Hands are a fundamental tool humans use to interact with the environment and objects. Through hand motions, we can obtain information about the shape and materials of the surfaces we touch, modify our surroundings by interacting with objects, manipulate objects and tools, or communicate with other people by leveraging the power of gestures. For these reasons, sensorized gloves, which can collect information about hand motions and interactions, have been of interest since the 1980s in various fields, such as Human-Machine Interaction (HMI) and the analysis and control of human motions. Over the last 40 years, research in this field explored different technological approaches and contributed to the popularity of wearable custom and commercial products targeting hand sensorization. Despite a positive research trend, these instruments are not widespread yet outside research environments and devices aimed at research are often ad hoc solutions with a low chance of being reused. This paper aims to provide a systematic literature review for custom gloves to analyze their main characteristics and critical issues, from the type and number of sensors to the limitations due to device encumbrance. The collection of this information lays the foundation for a standardization process necessary for future breakthroughs in this research field. © 2024, CC BY.

关键词： Man machine systems

Fabrication and characterization of plasmonic nanopores for Raman detection of biomolecules

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Iarossi, Marzia Department of Informatics Bioengineering Robotics and Systems Engineering Italy

In this thesis various plasmonic materials, such as hyperbolic metamaterial nanohole arrays and Au-coated free standing nanoassembly supporting nanopores, are investigated with the aim to perform optical sensing of DN... 详细信息

关键词： Nanopores

Development of a Novel Platform for in vitro Electrophysiological Recording

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Melle, Giovanni Department of Informatics Bioengineering Robotics and Systems Engineering Italy

The accurate monitoring of cell electrical activity is of fundamental importance for pharmaceutical research and pre-clinical trials that impose to check the cardiotoxicity of all new drugs. Traditional methods for preclinical evaluation of drug cardiotoxicity exploit animal models, which tend to be expensive, low throughput, and exhibit species-specific differences in cardiac physiology. Alternative approaches use heterologous expression of cardiac ion channels in non-cardiac cells transfected with genetic material. However, the use of these constructs and the inhibition of specific ionic currents alone is not predictive of cardiotoxicity. Drug toxicity evaluation based on the human ether-à-go-go-related gene (hERG) channel, for example, leads to a high rate of false-positive cardiotoxic compounds, increasing drug attrition at the preclinical stage. Consequently, from 2013, the Comprehensive in Vitro Proarrhythmia Assay (CiPA) initiative focused on experimental methods that identify cardiotoxic drugs and to improve upon prior models that have largely used alterations in the hERG potassium ion channel. The most predictive models for drug cardiotoxicity must recapitulate the complex spatial distribution of the physiologically distinct myocytes of the intact adult human heart. However, intact human heart preparations are inherently too costly, difficult to maintain, and, hence, too low throughput to be implemented early in the drug development pipeline. For these reasons the optimization of methodologies to differentiate human induced Pluripotent Stem Cells (hiPSCs) into cardiomyocytes (CMs) enabled human CMs to be mass-produced in vitro for cardiovascular disease modeling and drug screening 2. These hiPSC-CMs functionally express most of the ion channels and sarcomeric proteins found in adult human CMs and can spontaneously contract. Recent results from the CiPA initiative have confirmed that, if utilized appropriately, the hiPSC-CM platform can serve as a reliable a

关键词： Toxicity

Improving the ROS 2 Navigation Stack with Real-Time Local Costmap Updates for Agricultural Applications

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sani, Ettore Sgorbissa, Antonio Carpin, Stefano The Department of Informatics Bioengineering Robotics and Systems Engineering University of Genova Italy The Department of Computer Science and Engineering University of California MercedCA United States

The ROS 2 Navigation Stack (Nav2) has emerged as a widely used software component providing the underlying basis to develop a variety of high-level functionalities. However, when used in outdoor environments such as orchards and vineyards, its functionality is notably limited by the presence of obstacles and/or situations not commonly found in indoor settings. One such example is given by tall grass and weeds that can be safely traversed by a robot, but that can be perceived as obstacles by LiDAR sensors, and then force the robot to take longer paths to avoid them, or abort navigation altogether. To overcome these limitations, domain specific extensions must be developed and integrated into the software pipeline. This paper presents a new, lightweight approach to address this challenge and improve outdoor robot navigation. Leveraging the multiscale nature of the costmaps supporting Nav2, we developed a system that using a depth camera performs pixel level classification on the images, and in real time injects corrections into the local cost map, thus enabling the robot to traverse areas that would otherwise be avoided by the Nav2. Our approach has been implemented and validated on a Clearpath Husky and we demonstrate that with this extension the robot is able to perform navigation tasks that would be otherwise not practical with the standard components. © 2024, CC BY.

关键词： Air navigation

Locally Optimal Solutions to Constraint Displacement Problems Via Constraint Overlaps

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Thomas, Antony Mastrogiovanni, Fulvio Baglietto, Marco Department of Informatics Bioengineering Robotics and Systems Engineering University of Genoa Via All’Opera Pia 13 Genoa16145 Italy

We present a unified approach for constraint displacement problems in which a robot finds a feasible path by displacing constraints or obstacles. To this end, we propose a two stage process that returns locally optimal obstacle displacements to enable a feasible path for the robot. The first stage proceeds by computing a trajectory through the obstacles while minimizing an appropriate objective function. In the second stage, these obstacles are displaced to make the computed robot trajectory feasible (collision-free). Several examples are provided that successfully demonstrate our approach on two distinct classes of constraint displacement problems. © 2024, The Authors. All rights reserved.

关键词： Motion planning