國立高雄大學圖資館 |

登入

回首頁

Simple Models, Complex Worlds /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Simple Models, Complex Worlds /Veronika Grigoreva.
作者:	Grigoreva, Veronika,
面頁冊數:	1 electronic resource (76 pages)
附註:	Source: Masters Abstracts International, Volume: 87-06.
附註:	Advisors: Stinson, Catherine; Muise, Christian.
Contained By:	Masters Abstracts International87-06.
標題:	Skewness.
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=32353617
ISBN:	9798270206710

Simple Models, Complex Worlds /
Grigoreva, Veronika,

Simple Models, Complex Worlds /Veronika Grigoreva. - 1 electronic resource (76 pages)

Source: Masters Abstracts International, Volume: 87-06.

Using a simplified model of a real-life phenomenon instead of conducting experiments with the object of study itself can have multiple advantages, such as higher interpretability, lower resource requirements, and others. In this study, we explore two separate use cases of simplified computer models aimed at addressing two problems in different fields.In the first part of this study, we focus on large language models (LLMs). Modern language models train on vast arrays of often only lightly filtered data scraped from the Internet, and, considering the increasing amounts of machine-generated texts on the Internet, the possibility of a new language model training on the outputs of previous generations is exceedingly high. To analyse how repeated training on datasets that include AI-generated content might affect LLMs, we repeatedly simulate the process using a relatively small GPT-2 based LLM on several datasets. Afterwards, we score the outputs across multiple attributes, including quality, overall text diversity, as well as emotion, toxicity, perceived identity of the author, etc., using both established metrics and fine-tuned classifier models.As a second main exploration, we propose a minimal social model aimed at studying the conversation surrounding instances of sexual violence and the process of achieving social consensus in such cases from the epistemic perspective. This model accounts for several social factors, such as disparities in institutional power between agents, a victim's possible reluctance to come forward, etc., and simulates communication and opinion propagation between the agents. This design allows us to compare epistemic strategies and evaluate which produce more fair (unbiased) results.Lastly, we contrast the two models and their limitations and discuss key differences between modelling in these two fields and issues unique to each field. We find that both models provide meaningful results in their respective domains and are limited by their minimal design; however, the impact of a smaller size or a lower complexity differs between fields.

English

ISBN: 9798270206710Subjects--Topical Terms:

1010492
Skewness.

Simple Models, Complex Worlds /
LDR:03458nmm a22004453i 4500 001 693797
005 20260511063026.5
006 m o d
007 cr|nu||||||||
008 260819s2025 miu||||||m |||||||eng d
020 $a 9798270206710
035 $a (MiAaPQD)AAI32353617
035 $a AAI32353617
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Grigoreva, Veronika, $e author. $3 1010491
245 1 0 $a Simple Models, Complex Worlds / $c Veronika Grigoreva.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2025
300 $a 1 electronic resource (76 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 87-06.
500 $a Advisors: Stinson, Catherine; Muise, Christian.
502 $b M.S. $c Queen's University (Canada) $d 2025.
520 $a Using a simplified model of a real-life phenomenon instead of conducting experiments with the object of study itself can have multiple advantages, such as higher interpretability, lower resource requirements, and others. In this study, we explore two separate use cases of simplified computer models aimed at addressing two problems in different fields.In the first part of this study, we focus on large language models (LLMs). Modern language models train on vast arrays of often only lightly filtered data scraped from the Internet, and, considering the increasing amounts of machine-generated texts on the Internet, the possibility of a new language model training on the outputs of previous generations is exceedingly high. To analyse how repeated training on datasets that include AI-generated content might affect LLMs, we repeatedly simulate the process using a relatively small GPT-2 based LLM on several datasets. Afterwards, we score the outputs across multiple attributes, including quality, overall text diversity, as well as emotion, toxicity, perceived identity of the author, etc., using both established metrics and fine-tuned classifier models.As a second main exploration, we propose a minimal social model aimed at studying the conversation surrounding instances of sexual violence and the process of achieving social consensus in such cases from the epistemic perspective. This model accounts for several social factors, such as disparities in institutional power between agents, a victim's possible reluctance to come forward, etc., and simulates communication and opinion propagation between the agents. This design allows us to compare epistemic strategies and evaluate which produce more fair (unbiased) results.Lastly, we contrast the two models and their limitations and discuss key differences between modelling in these two fields and issues unique to each field. We find that both models provide meaningful results in their respective domains and are limited by their minimal design; however, the impact of a smaller size or a lower complexity differs between fields.
546 $a English
590 $a School code: 0283
650 4 $a Skewness. $3 1010492
650 4 $a Toxicity. $3 967024
650 4 $a Haiku. $3 235088
650 4 $a Logic. $3 180785
650 4 $a Gender. $3 915601
650 4 $a Design. $3 204000
650 4 $a Linguistics. $3 176897
650 4 $a Sex crimes. $3 327571
650 4 $a Realism. $3 174751
650 4 $a Large language models. $3 1010493
650 4 $a Chatbots. $3 970198
650 4 $a Cognition & reasoning. $3 1010494
650 4 $a Natural language. $3 1010495
650 4 $a Cognitive psychology. $3 179548
650 4 $a Criminology. $3 190212
650 4 $a Web studies. $3 708690
650 4 $a Toxicology. $3 206218
690 $a 0290
690 $a 0395
690 $a 0389
690 $a 0800
690 $a 0633
690 $a 0627
690 $a 0646
690 $a 0383
710 2 $a Queen's University (Canada). $3 766172
720 1 $a Stinson, Catherine $e degree supervisor.
720 1 $a Muise, Christian $e degree supervisor.
773 0 $t Masters Abstracts International $g 87-06.
790 $a 0283
791 $a M.S.
792 $a 2025
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=32353617