@inproceedings{DPSP22,

title = {DeTox: A Comprehensive Dataset for German Offensive Language and Conversation Analysis},

author = {Christoph Demus and Jonas Pitz and Mina Schütz and Nadine Probol and Melanie Siegel and Dirk Labudde},

url = {https://aclanthology.org/2022.woah-1.14},

doi = {10.18653/v1/2022.woah-1.14},

year  = {2022},

date = {2022-07-01},

booktitle = {Proceedings of the Sixth Workshop on Online Abuse and Harms (WOAH)},

pages = {143--153},

publisher = {Association for Computational Linguistics},

address = {Seattle, Washington (Hybrid)},

abstract = {In this work, we present a new publicly available offensive language dataset of 10.278 German social media comments collected in the first half of 2021 that were annotated by in total six annotators. With twelve different annotation categories, it is far more comprehensive than other datasets, and goes beyond just hate speech detection. The labels aim in particular also at toxicity, criminal relevance and discrimination types of comments.Furthermore, about half of the comments are from coherent parts of conversations, which opens the possibility to consider the comments' contexts and do conversation analyses in order to research the contagion of offensive language in conversations.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{SiNa22,

title = {Visual Analytics for Systematic Reviews According to PRISMA},

author = {Lennart B. Sina and Kawa Nazemi},

doi = {10.1109/IV56949.2022.00059},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

booktitle = {2022 26th International Conference Information Visualisation (IV)},

pages = {307 - 313},

publisher = {IEEE},

abstract = {Systematic reviews play an essential role in various disciplines. Particularly, in biomedical sciences, systematic reviews according to a predefined schema and protocol are how related literature is analyzed. Although a protocol-based systematic review is replicable and provides the required information to reproduce each step and refine them, such a systematic review is time-consuming and may get complex. To face this challenge, automatic methods can be applied that support researchers in their systematic analysis process. The combination of artificial intelligence for automatic information extraction from scientific literature with interactive visualizations as a Visual Analytics system can lead to sophisticated analysis and protocoling of the review process. We introduce in this paper a novel Visual Analytics approach and system that enables researchers to visually search and explore scientific publications and generate a protocol based on the PRISMA protocol and the PRISMA statement.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{BSN22,

title = { Visual Collaboration - An Approach for Visual Analytical Collaborative Research},

author = {Midhad Blazevic and Lennart B. Sina and Kawa Nazemi},

doi = {10.1109/IV56949.2022.00057},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

booktitle = {2022 26th International Conference Information Visualisation (IV)},

pages = {293  - 299},

publisher = {IEEE},

abstract = {Studies have shown that collaboration in scientific fields is rising and considered enormously important. However, collaboration has proved to be challenging for various reasons, among others, the requirements for human-machine workflows. The importance of scientific collaboration lies in the complexity of the challenges that are faced today. The more complex the challenge, the more scientists should work together. The current form of collaboration in the scientific community is not as intelligent as it should be. Scientists have to multitask with various applications, often losing cognitive focus. Collaboration itself is very nearsighted as it is usually conducted not solely based on expertise but instead on social or local networks. We introduce a single-source visual collaboration approach based on learning methods in this work. We use machine learning and natural language processing approaches to improve the traditional research and development process and create a system that facilitates and encourages collaboration based on expertise, enhancing the research collaboration process in many ways. Our approach combines collaborative Visual Analytics with enhanced collaboration techniques to support researchers from different disciplines.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{KSSW22,

title = {Overview of the CLEF-2022 CheckThat! Lab: Task 3 on Fake News Detection},

author = {Juliane Köhler and Gautam Kishore Shahi and Julia Maria Struss and Michael Wiegand and Melanie Siegel and Mina Schütz},

editor = {Guglielmo Faggioli and Nicola Ferro and Allan Hanbury and Martin Potthast},

url = {https://ceur-ws.org/Vol-3180/paper-30.pdf},

year  = {2022},

date = {2022-01-01},

booktitle = {Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum},

pages = {404-421},

address = {Bologna, Italy},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{SBAK22,

title = {AIT_FHSTP at CheckThat! 2022: Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer},

author = {Mina Schütz and Jaqueline Böck and Medina Andresel and Armin Kirchknopf and Daria Liakhovets and Djordje Slijepčević and Alexander Schindler},

editor = {Guglielmo Faggioli and Nicola Ferro and Allan Hanbury and Martin Potthast},

url = {https://ceur-ws.org/Vol-3180/paper-53.pdf},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

booktitle = {Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum},

pages = {660-670},

address = {Bologna, Italy},

abstract = {The increase of fake news in today’s society, partially due to the accelerating digital transformation, is a major problem in today’s world. This year’s CheckThat! Lab 2022 challenge addresses this problem as a Natural Language Processing (NLP) task aiming to detect fake news in English and German texts. Within this paper, we present our methodology and results for both, the monolingual (English) and cross-lingual (German) tasks of the CheckThat! challenge in 2022. We applied the multilingual transformer model XLM-RoBERTa to solve these tasks by pre-training the models on additional datasets and fine-tuning them on the original data as well as its translations for the cross-lingual task. Our final model achieves a macro F1-score of 15,48% and scores the 22𝑡ℎ rank in the benchmark. Regarding the second task, i.e., the cross-lingual German classification, our final model achieves an F1-score of 19.46% and reaches the 4𝑡ℎ rank in the benchmark.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{LSBA22,

title = {Transfer Learning for Automatic Sexism Detection with Multilingual Transformer Models},

author = {Daria Liakhovets and Mina Schütz and Jaqueline Böck and Medina Andresel and Armin Kirchknopf and Andreas Babic and Djordje Slijepčević and Jasmin Lampert and Alexander Schindler and Matthias Zeppelzauer},

editor = {Manuel Montes-y-Gómez and Julio Gonzalo and Francisco Rangel and Marco Casavantes and Miguel Ángel Álvarez-Carmona and Gemma Bel-Enguix and Hugo Jair Escalante and Larissa Freitas and Antonio Miranda-Escalada and Francisco Rodríguez-Sánchez and Aiala Rosá and Marco Antonio Sobrevilla-Cabezudo and Mariona Taulé and Rafael Valencia-García},

url = {https://ceur-ws.org/Vol-3202/exist-paper1.pdf},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

booktitle = {Proceedings of the Iberian Languages Evaluation Forum (IberLEF 2022)},

publisher = {CEUR-WS.org},

address = {A Coruna, Spain},

abstract = {In recent years sexism has become an increasingly significant problem on social networks. In order to address this problem, the sEXism Identification in Social neTworks (EXIST) challenge has been launched at IberLEF in 2021. In this international benchmark, sexism detection is formulated as a Natural Language Processing (NLP) task with the aim to automatically identify sexism in social media content (binary classification) and to classify statements into different categories such as dominance, stereotyping or objectification. In this paper we present the contribution of team AIT_FHSTP for the EXIST challenge at IberLEF in 2022. To solve the two related tasks we applied two multilingual transformer models, one based on a multilingual BERT and one based on an XLM-RoBERTa architecture, and a monolingual (English) T5 model. Our approach uses two different strategies to adapt the transformers to the detection of sexist content: first, unsupervised pre-training with additional data and second, supervised fine-tuning with additional as well as augmented data. For both tasks the XLM-RoBERTa model, which applies a combination of the two strategies, outperforms the other two models. The best run for the binary classification (task 1) achieves a macro F1-score of 74.96% and scores the 26𝑡ℎ rank in the benchmark; for the multi-class classification (task 2) our best submission scores the 13𝑡ℎ rank with a macro F1-score of 46.75%.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{PSSL22,

title = {Deep Learning Frameworks Applied For Audio-Visual Scene Classification},

author = {Lam Pham and Alexander Schindler and Mina Schütz and Jasmin Lampert and Sven Schlarb and Ross King},

editor = {Peter Haber and Thomas J. Lampoltshammer and Helmut Leopold and Manfred Mayr},

isbn = {978-3-658-36295-9},

year  = {2022},

date = {2022-01-01},

booktitle = {Data Science -- Analytics and Applications},

pages = {39--44},

publisher = {Springer Fachmedien Wiesbaden},

address = {Wiesbaden},

abstract = {In this paper, we present deep learning frameworks for audio-visual scene classification (SC) and indicate how individual visual, audio features as well as their combination affect SC performance. Our extensive experiments are conducted on DCASE 2021 (IEEE AASP Challenge on Detection and Classification of Acoustic Scenes and Events) Task 1B Development and Evaluation datasets. Our results on Development dataset achieve the best classification accuracy of 82.2%, 91.1%, and 93.9% with audio input only, visual input only, and both audio-visual input, respectively. The highest classification accuracy of 93.9%, obtained from an ensemble of audio-based and visual-based frameworks, shows an improvement of 16.5% compared with DCASE 2021 baseline. Our best results on Evaluation dataset is 91.5%, outperforming DCASE baseline of 77.1%},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{ILSN22,

title = {Combining social media open source data with relevance analysis and expert knowledge to improve situational awareness in crisis and disaster management - concept},

doi = {10.35011/IDIMT-2022-153},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

booktitle = {IDIMT-2022 : digitalization of society, business and management in a pandemic : 30th Interdisciplinary Information Management Talks},

publisher = {Linz Trauner Verlag},

abstract = {Situational awareness is one of the most important factors for efficient and effective response in crisis and disaster situations. Up-to-date, valid and relevant data is one of the means to support crisis management actions, and the development and use of social media, as it is common nowadays, has become a very interesting research topic. In this paper, we describe a concept for using social media information and open source data in combination with expert knowledge and relevance assessment. The main problem with data collected from these sources is, on the one hand, the large amount and difficulty of processing, and on the other hand, the difficulty of determining its validity. This concept aims to address these two problems and combine the results to propose a solution for better situational awareness in crisis and disaster management. Situational awareness is one of the most important factors for efficient and effective response in crisis and disaster situations. Up-to-date, valid and relevant data is one of the means to support crisis management actions, and the development and use of social media, as it is common nowadays, has become a very interesting research topic. In this paper, we describe a concept for using social media information and open source data in combination with expert knowledge and relevance assessment. The main problem with data collected from these sources is, on the one hand, the large amount and difficulty of processing, and on the other hand, the difficulty of determining its validity. This concept aims to address these two problems and combine the results to propose a solution for better situational awareness in crisis and disaster management.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{BLSK21,

title = {AIT_FHSTP at GermEval 2021: Automatic Fact Claiming Detection with Multilingual Transformer Models},

author = {Jaqueline Böck and Daria Liakhovets and Mina Schütz and Armin Kirchknopf and Djordje Slijepčevi'c and Matthias Zeppelzauer and Alexander Schindler},

url = {https://aclanthology.org/2021.germeval-1.11},

year  = {2021},

date = {2021-09-01},

booktitle = {Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments},

pages = {76--82},

publisher = {Association for Computational Linguistics},

address = {Duesseldorf, Germany},

abstract = {Spreading ones opinion on the internet is becoming more and more important. A problem is that in many discussions people often argue with supposed facts. This year's GermEval 2021 focuses on this topic by incorporating a shared task on the identification of fact-claiming comments. This paper presents the contribution of the AIT FHSTP team at the GermEval 2021 benchmark for task 3: ``identifying fact-claiming comments in social media texts''. Our methodological approaches are based on transformers and incorporate 3 different models: multilingual BERT, GottBERT and XML-RoBERTa. To solve the fact claiming task, we fine-tuned these transformers with external data and the data provided by the GermEval task organizers. Our multilingual BERT model achieved a precision-score of 72.71%, a recall of 72.96% and an F1-Score of 72.84% on the GermEval test set. Our fine-tuned XML-RoBERTa model achieved a precision-score of 68.45%, a recall of 70.11% and a F1-Score of 69.27%. Our best model is GottBERT (i.e., a BERT transformer pre-trained on German texts) fine-tuned on the GermEval 2021 data. This transformer achieved a precision of 74.13%, a recall of 75.11% and an F1-Score of 74.62% on the test set.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{SDPP21,

title = {DeTox at GermEval 2021: Toxic Comment Classification},

author = {Mina Schütz and Christoph Demus and Jonas Pitz and Nadine Probol and Melanie Siegel and Dirk Labudde},

url = {https://aclanthology.org/2021.germeval-1.8},

year  = {2021},

date = {2021-09-01},

booktitle = {Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments},

pages = {54--61},

publisher = {Association for Computational Linguistics},

address = {Duesseldorf, Germany},

abstract = {In this work, we present our approaches on the toxic comment classification task (subtask 1) of the GermEval 2021 Shared Task. For this binary task, we propose three models: a German BERT transformer model; a multilayer perceptron, which was first trained in parallel on textual input and 14 additional linguistic features and then concatenated in an additional layer; and a multilayer perceptron with both feature types as input. We enhanced our pre-trained transformer model by re-training it with over 1 million tweets and fine-tuned it on two additional German datasets of similar tasks. The embeddings of the final fine-tuned German BERT were taken as the textual input features for our neural networks. Our best models on the validation data were both neural networks, however our enhanced German BERT gained with a F1-score = 0.5895 a higher prediction on the test data.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{10.1007/978-3-030-68787-8_45,

title = {Automatic Fake News Detection with Pre-trained Transformer Models},

author = {Mina Schütz and Alexander Schindler and Melanie Siegel and Kawa Nazemi},

editor = {Alberto Del Bimbo and Rita Cucchiara and Stan Sclaroff and Giovanni Maria Farinella and Tao Mei and Marco Bertini and Hugo Jair Escalante and Roberto Vezzani},

doi = {10.1007/978-3-030-68787-8_45},

isbn = {978-3-030-68787-8},

year  = {2021},

date = {2021-02-21},

urldate = {2021-02-21},

booktitle = {Pattern Recognition. ICPR International Workshops and Challenges},

pages = {627--641},

publisher = {Springer International Publishing},

address = {Cham},

abstract = {The automatic detection of disinformation and misinformation has gained attention during the last years, since fake news has a critical impact on democracy, society, and journalism and digital literacy. In this paper, we present a binary content-based classification approach for detecting fake news automatically, with several recently published pre-trained language models based on the Transformer architecture. The experiments were conducted on the FakeNewsNet dataset with XLNet, BERT, RoBERTa, DistilBERT, and ALBERT and various combinations of hyperparameters. Different preprocessing steps were carried out with only using the body text, the titles and a concatenation of both. It is concluded that Transformers are a promising approach to detect fake news, since they achieve notable results, even without using a large dataset. Our main contribution is the enhancement of fake news' detection accuracy through different models and parametrizations with a reproducible result examination through the conducted experiments. The evaluation shows that already short texts are enough to attain 85% accuracy on the test set. Using the body text and a concatenation of both reach up to 87% accuracy. Lastly, we show that various preprocessing steps, such as removing outliers, do not have a significant impact on the models prediction output.},

keywords = {Artificial Intelligence, Data Analytics, Data Mining, Fake News, maschine learning, Transformer},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{Kaupp2021,

title = {CONTEXT: An Industry 4.0 Dataset of Contextual Faults in a Smart Factory},

author = {Lukas Kaupp and Heiko Webert and Kawa Nazemi and Bernhard Humm and Stephan Simons},

doi = {10.1016/j.procs.2021.01.265},

issn = {1877-0509},

year  = {2021},

date = {2021-02-17},

booktitle = {Proceedings of the 2nd International Conference on Industry 4.0 and Smart Manufacturing (ISM 2020)},

volume = {180},

pages = {492-501},

publisher = {Elsevier},

series = {Procedia Computer Science},

abstract = {Cyber-physical systems in smart factories get more and more integrated and interconnected. Industry 4.0 accelerates this trend even further. Through the broad interconnectivity a new class of faults arise, the contextual faults, where contextual knowledge is needed to find the underlying reason. Fully-automated systems and the production line in a smart factory form a complex environment making the fault diagnosis non-trivial. Along with a dataset, we give a first definition of contextual faults in the smart factory and name initial use cases. Additionally, the dataset encompasses all the data recorded in a current state-of-the-art smart factory. We also add additional information measured by our developed sensing units to enrich the smart factory data even further. In the end, we show a first approach to detect the contextual faults in a manual preliminary analysis of the recorded log data.},

keywords = {anomaly detection, contextual faults, cyber-physical systems, fault diagnosis, smart factory},

pubstate = {published},

tppubtype = {inproceedings}

}