--- _id: '34025' abstract: - lang: eng text: "Controversial topics like abortion or capital punishment inherently lack\r\nof correct answers or the right way to deal with. Thus, in order to find what is true,\r\nwhat is good, or what should be done, the involved parties need to debate. For the\r\npurpose of forming an opinion on a controversial topic someone needs to take in a\r\nlot of arguments on that topic to gather information which can be a time-consuming\r\nprocess. To increase efficiency, someone can use an argument search engine to quicken\r\nthe retrieval of relevant arguments. Although the usage of such a service reduces the\r\ntime to find arguments, there is still a lot of textual data that needs to be read. To this\r\nend, computational summarization approaches for arguments can limit the necessary\r\ntime for information review by generating short snippets capturing the main gist of\r\neach argument. Yet, we suggest that approaches that consider one argument at a\r\ntime show potential for further improvement in terms of efficiency during information\r\nreview. In fact, arguments on the same topic, like those retrieved by a search engine for\r\na certain query, partially cover the same content, e. g. arguments regarding the death\r\npenalty probably use deterrence as a point in favor of it. However, if the same aspect\r\nis central in multiple arguments, their snippets reflect this, which leads to redundancy\r\namong the snippets. Consequently, someone interested in gathering information on a\r\ncontroversial topic does not necessarily find new information in each snippet he or she\r\nreads.\r\nWe introduce the task of Contrastive Argument Summarization (CAS) which addresses\r\nthe aforementioned problem regarding existing argument summarization. An approach\r\nthat addresses CAS aims to produce contrastive snippets for each argument in a set\r\nof topic-related arguments. A contrastive snippet should represent the main gist of its\r\nargument, it should account for the argumentative nature of the text, and it should be\r\ndissimilar to the other topic-related arguments in order to reduce redundancy among\r\nthe snippets.\r\nWe propose two approaches addressing CAS, namely an extended version of the\r\nLexRank derivation by Alshomary et al. (2020), and an advancement of the work\r\nby Bista et al. (2020). Additionally, we develop two automatic measures to assess to\r\nwhich extent the snippets of one set are opposed. For evaluation, we compile a corpus\r\nusing the args.me search engine Wachsmuth et al. (2017b) to come close to the suggested area of application. Moreover, we conduct a manual annotation study to assess\r\napproaches’ effectiveness. We find that the graph-based approach is superior when it\r\ncomes to contrastiveness (i. e. snippets being dissimilar to topic-related arguments),\r\nand that the second approach outperforms the previous one and the unmodified version of Alshomary et al. (2020) when it comes to representativeness (i. e. snippets\r\ncapturing the main gist of an argument)." author: - first_name: Jonas full_name: Rieskamp, Jonas id: '77643' last_name: Rieskamp citation: ama: Rieskamp J. Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning.; 2022. apa: Rieskamp, J. (2022). Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning. bibtex: '@book{Rieskamp_2022, title={Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning}, author={Rieskamp, Jonas}, year={2022} }' chicago: Rieskamp, Jonas. Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning, 2022. ieee: J. Rieskamp, Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning. 2022. mla: Rieskamp, Jonas. Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning. 2022. short: J. Rieskamp, Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning, 2022. date_created: 2022-11-07T13:57:08Z date_updated: 2022-11-07T13:57:37Z language: - iso: eng main_file_link: - url: https://en.cs.uni-paderborn.de/fileadmin/informatik/fg/css/teaching/theses/thesis_final.pdf status: public supervisor: - first_name: Milad full_name: Alshomary, Milad id: '73059' last_name: Alshomary - first_name: Henning full_name: Wachsmuth, Henning id: '3900' last_name: Wachsmuth title: Contrastive Argument Summarization Using Supervised and Unsupervised Machine Learning type: mastersthesis user_id: '77643' year: '2022' ...