检索结果-内蒙古大学图书馆

Fragile, yet resilient: Adaptive decline in a collaboration network of firms

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Schweitzer, Frank Casiraghi, Giona Tomasello, Mario V. Garcia, David ETH Zurich Switzerland Complexity Science Hub Vienna Austria Institute of Interactive Systems and Data Science Graz University of Technology Austria

The dynamics of collaboration networks of firms follow a life-cycle of growth and decline. That does not imply they also become less resilient. Instead, declining collaboration networks may still have the ability to mitigate shocks from firms leaving, and to recover from these losses by adapting to new partners. To demonstrate this, we analyze 21.500 R&D collaborations of 14.500 firms in six different industrial sectors over 25 years. We calculate time-dependent probabilities of firms leaving the network and simulate drop-out cascades, to determine the expected dynamics of decline. We then show that deviations from these expectations result from the adaptivity of the network, which mitigates the decline. These deviations can be used as a measure of network resilience. Copyright © 2020, The Authors. All rights reserved.

关键词：

Detecting potentially harmful and protective suicide-related content on twitter: A machine learning approach

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Metzler, Hannah Baginski, Hubert Niederkrotenthaler, Thomas Garcia, David Section for Science of Complex Systems Center for Medical Statistics Informatics and Intelligent Systems Medical University of Vienna Austria Unit Suicide Research & Mental Health Promotion Department of Social and Preventive Medicine Center for Public Health Medical University of Vienna Austria Complexity Science Hub Vienna Austria Institute of Interactive Systems and Data Science Department of Computer Science and Biomedical Engineering Graz University of Technology Graz Austria Institute of Globally Distributed Open Research and Education Austria Institute of Information Systems Engineering Vienna University of Technology Vienna Austria

Background Research has repeatedly shown that exposure to suicide-related news media content is associated with suicide rates, with some content characteristics likely having harmful and others potentially protective effects. Although good evidence exists for a few selected characteristics, systematic and large-scale investigations are missing. Moreover, the growing importance of social media, particularly among young adults, calls for studies on the effects of content posted on these platforms. Objective This study applies natural language processing and machine learning methods to classify large quantities of social media data according to characteristics identified as potentially harmful or beneficial in media effects research on suicide and prevention. Methods We manually labeled 3202 English tweets using a novel annotation scheme that classifies suicide-related tweets into 12 categories. Based on these categories, we trained a benchmark of machine learning models for a multi-class and a binary classification task. As models, we included a majority classifier, an approach based on word frequency (TF-IDF with a linear SVM) and two state-of-the-art deep learning models (BERT, XLNet). The first task classified postings into six main content categories, which are particularly relevant for suicide prevention based on previous evidence. These included personal stories of either (1) suicidal ideation and attempts or (2) coping and recovery, calls for action intending to spread either (3) problem awareness or (4) prevention-related information, (5) reporting of suicide cases, and (6) other tweets irrelevant to these five categories. The second classification task was binary, and separated postings in the 11 categories that refer to actual suicide, from postings in the off-topic category, which use suicide-related terms in another meaning or context. Results In both tasks, the performance of the two deep learning models was very similar and better than the majority or th

关键词： Learning algorithms

Vehicle management in a modular production context using Deep Q-Learning

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Pouget, Lucain Hasenbichler, Timo Auer, Jakob Lichtenegger, Klaus Windisch, Andreas Skalar Systems GmbH Dr.-Robert-Graf-Straße 9 Graz8010 Austria FH JOANNEUM University of Applied Sciences Data Science and Artificial Intelligence Eckertstraße 30i Graz8020 Austria Know-Center GmbH Inffeldgasse 13 Graz8010 Austria Institute of Interactive Systems and Data Science Graz University of Technology Inffeldgasse 13 Graz8010 Austria Physics Department Washington University in St. Louis One Brookings Drive St. LouisMO63130 United States RL Community AI Austria Wollzeile 24/12 Vienna1010 Austria

We investigate the feasibility of deploying Deep-Q based deep reinforcement learning agents to job-shop scheduling problems in the context of modular production facilities, using discrete event simulations for the environment. These environments are comprised of a source and sink for the parts to be processed, as well as (several) workstations. The agents are trained to schedule automated guided vehicles to transport the parts back and forth between those stations in an optimal fashion. Starting from a very simplistic setup, we increase the complexity of the environment and compare the agents' performances with well established heuristic approaches, such as first-in-first-out based agents, cost tables and a nearest-neighbor approach. We furthermore seek particular configurations of the environments in which the heuristic approaches struggle, to investigate to what degree the Deep-Q agents are affected by these challenges. We find that Deep-Q based agents show comparable performance as the heuristic baselines. Furthermore, our findings suggest that the DRL agents exhibit an increased robustness to noise, as compared to the conventional approaches. Overall, we find that DRL agents constitute a valuable approach for this type of scheduling problems. Copyright © 2022, The Authors. All rights reserved.

关键词： Deep learning

Social media emotion macroscopes reflect emotional experiences in society at large

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Garcia, David Pellert, Max Lasser, Jana Metzler, Hannah Graz University of Technology Institute of Interactive Systems and Data Science Department of Computer Science and Biomedical Engineering Graz8010 Austria Complexity Science Hub Vienna Vienna1080 Austria Medical University of Vienna Center for Medical Statistics Informatics and Intelligent Systems Vienna1090 Austria Medical University of Vienna Department of Social and Preventive Medicine Center for Public Health Vienna1090 Austria

Social media generate data on human behaviour at large scales and over long periods of time, posing a complementary approach to traditional methods in the social sciences1. Millions of texts from social media can be processed with computational methods to study emotions2,3 over time and across regions4,5. However, recent research has shown weak correlations between social media emotions and affect questionnaires at the individual level6,7 and between static regional aggregates of social media emotion and subjective well-being at the population level8, questioning the validity of social media data to study emotions. Yet, to date, no research has tested the validity of social media emotion macroscopes to track the temporal evolution of emotions at the level of a whole society. Here we present a pre-registered prediction study that shows how gender-rescaled time series of Twitter emotional expression at the national level substantially correlate with aggregates of self-reported emotions in a weekly representative survey in the United Kingdom. A follow-up exploratory analysis shows a high prevalence of third-person references in emotionally-charged tweets, indicating that social media data provide a way of social sensing the emotions of others9 rather than just the emotional experiences of users. These results show that, despite the issues that social media have in terms of representativeness10 and algorithmic confounding11, the combination of advanced text analysis methods with user demographic information in social media emotion macroscopes can provide measures that are informative of the general population beyond social media users. Copyright © 2021, The Authors. All rights reserved.

关键词： Social networking (online)

Teaching mechanics with individual exercise assignments and automated correction

学校读者我要写书评

暂无评论

PAMM 2023年第3期23卷

作者： Michael H. Gfrerer Benjamin Marussig Katharina Maitz Mia M. Bangerl Institute of Applied Mechanics Graz Center of Computational Engineering (GCCE) Graz University of Technology Graz Austria Private University College of Teacher Education Augustinum (PPH Augustinum) Graz Austria Institute of Interactive Systems and Data Science Graz University of Technology Graz Austria

Solving exercise problems by yourself is a vital part of developing a mechanical understanding. Yet, most mechanics lectures have more than 200 participants, so the workload for manually creating and correcting assignments limits the number of exercises. The resulting example pool is usually much smaller than the number of participants, making verifying whether students can solve problems themselves considerably harder. At the same time, unreflected copying of tasks already solved does not foster the understanding of the subject and leads to a false self-assessment. We address these issues by providing a scalable approach for creating, distributing, and correcting exercise assignments for problems related to statics, strength of materials, dynamics, and hydrostatics. The overall concept allows us to provide individual exercise assignments for each student. A quantitative survey among students of our recent statics lecture assesses the acceptance of our teaching tool. The feedback indicates a clear added value for the lecture, which fosters self-directed and reflective learning.

关键词：

Evaluation of data Augmentation and Loss Functions in Semantic Image Segmentation for Drilling Tool Wear Detection

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Schlager, Elke Windisch, Andreas Hanna, Lukas Klünsner, Thomas Hagendorfer, Elias Jan Teppernegg, Tamara Know-Center GmbH Sandgasse 36 Graz8010 Austria Graz University of Technology Institute of Interactive Systems and Data Science Sandgasse 36 Graz8010 Austria Washington University in St. Louis Physics Department One Brookings Drive St. LouisMO63130 United States Reinforcement Learning Community AI Austria Wollzeile 24/12 Vienna1010 Austria Materials Center Leoben Forschung GmbH Roseggerstrasse 12 Leoben8700 Austria CERATIZIT Austria GmbH Metallwerk-Plansee-Straße 71 Breitenwang6600 Austria

Tool wear monitoring is crucial for quality control and cost reduction in manufacturing processes, of which drilling applications are one example. In this paper, we present a U-Net based semantic image segmentation pipeline, deployed on microscopy images of cutting inserts, for the purpose of wear detection. The wear area is differentiated in two different types, resulting in a multiclass classification problem. Joining the two wear types in one general wear class, on the other hand, allows the problem to be formulated as a binary classification task. Apart from the comparison of the binary and multiclass problem, also different loss functions, i. e., Cross Entropy, Focal Cross Entropy, and a loss based on the Intersection over Union (IoU), are investigated. Furthermore, models are trained on image tiles of different sizes, and augmentation techniques of varying intensities are deployed. We find, that the best performing models are binary models, trained on data with moderate augmentation and an IoU-based loss function. Copyright © 2023, The Authors. All rights reserved.

关键词： Wear of materials

interactive visual labelling versus active learning:an experimental comparison

学校读者我要写书评

暂无评论

Frontiers of Information Technology & Electronic Engineering 2020年第4期21卷 524-535页

作者： Mohammad CHEGINI Jurgen BERNARD Jian CUI Fatemeh CHEGINI Alexei SOURIN Keith ANDREWS Tobias SCHRECK Institute of Computer Graphics and Knowledge Visualisation Graz University of TechnologyGraz 8010Austria School of Computer Science and Engineering Nanyang Technological UniversitySingapore 639798Singapore InfoVis Group University of British ColumbiaVancouver V6T1Z4Canada Max Planck Institute for Meteorology Hamburg 20146Germany Institute of Interactive Systems and Data Science Graz University of TechnologyGraz 8010Austria

Methods from supervised machine learning allow the classification of new data automatically and are tremendously helpful for data *** quality of supervised maching learning depends not only on the type of algorithm used,but also on the quality of the labelled dataset used to train the *** instances in a training dataset is often done manually relying on selections and annotations by expert analysts,and is often a tedious and time-consuming *** learning algorithms can automatically determine a subset of data instances for which labels would provide useful input to the learning *** visual labelling techniques are a promising alternative,providing effective visual overviews from which an analyst can simultaneously explore data records and select items to a *** putting the analyst in the loop,higher accuracy can be achieved in the resulting *** initial results of interactive visual labelling techniques are promising in the sense that user labelling can improve supervised learning,many aspects of these techniques are still largely *** paper presents a study conducted using the mVis tool to compare three interactive visualisations,similarity map,scatterplot matrix(SPLOM),and parallel coordinates,with each other and with active learning for the purpose of labelling a multivariate *** results show that all three interactive visual labelling techniques surpass active learning algorithms in terms of classifier accuracy,and that users subjectively prefer the similarity map over SPLOM and parallel coordinates for *** also employ different labelling strategies depending on the visualisation used.

关键词： interactive visual labelling Active learning Visual analytics

On the Role of Priors in Bayesian Causal Learning

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Geiger, Bernhard C. Kern, Roman Signal Processing and Speech Communication Laboratory Graz University of Technology Inf-feldgasse 16c Graz8010 Austria Know Center Research GmbH Sandgasse 34 Graz8010 Austria Institute for Interactive Systems and Data Science Graz University of Technology Sandgasse 36 Graz8010 Austria

In this work, we investigate causal learning of independent causal mechanisms from a Bayesian perspective. Confirming previous claims from the literature, we show in a didactically accessible manner that unlabeled data (i.e., cause realizations) do not improve the estimation of the parameters defining the mechanism. Furthermore, we observe the importance of choosing an appropriate prior for the cause and mechanism parameters, respectively. Specifically, we show that a factorized prior results in a factorized posterior, which resonates with Janzing and Schölkopf’s definition of independent causal mechanisms via the Kolmogorov complexity of the involved distributions and with the concept of parameter independence of Heckerman et al. Copyright © 2025, The Authors. All rights reserved.

关键词：

Validating daily social media macroscopes of emotions

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Pellert, Max Metzler, Hannah Matzenberger, Michael Garcia, David Sony Computer Science Laboratories 6 Rue Amyot Paris75005 France Institute of Interactive Systems and Data Science Graz University of Technology Inffeldgasse 16C Graz8010 Austria Complexity Science Hub Vienna Josefstädter Straße 39 Vienna1080 Austria Section for Science of Complex Systems Center for Medical Statistics Informatics and Intelligent Systems Medical University of Vienna Spitalgasse 23 Vienna1090 Austria Institute of Globally Distributed Open Research and Education Sweden Der Standard Vordere Zollamtsstraße 13 Vienna1030 Austria

Measuring sentiment in social media text has become an important practice in studying emotions at the macroscopic level. However, this approach can suffer from methodological issues like sampling biases and measurement errors. To date, it has not been validated if social media sentiment can actually measure the temporal dynamics of mood and emotions aggregated at the level of communities. We ran a large-scale survey at an online newspaper to gather daily mood self-reports from its users, and compare these with aggregated results of sentiment analysis of user discussions. We find strong correlations between text analysis results and levels of self-reported feelings, as well as between inter-day changes of both measurements. We replicate these results using sentiment data from Twitter. We show that a combination of supervised text analysis methods based on novel deep learning architectures and unsupervised dictionary-based methods have high agreement with the time series of aggregated mood measured with self-reports. Our findings indicate that macro level dynamics of feelings expressed on an online platform can be tracked with social media text, especially in situations of high mood variability. © 2021, CC BY.

关键词： Sentiment analysis