Please use this identifier to cite or link to this item:
http://earchive.tpu.ru/handle/11683/77924| Title: | Problem of cheating with neural networks in language learning |
| Authors: | Зайда, А. В. Согуляк, Я. А. |
| metadata.dc.contributor.advisor: | Аксёнова, Наталия Валерьевна |
| Keywords: | электронный ресурс; труды учёных ТПУ; language learning; cheating; large language models; task evaluation; task adaptation |
| Issue Date: | 2023 |
| Publisher: | Томский политехнический университет |
| Citation: | Зайда, А. В. Problem of cheating with neural networks in language learning / А. В. Зайда, Я. А. Согуляк ; науч. рук. Н. В. Аксёнова ; Национальный исследовательский Томский политехнический университет // Язык. Общество. Образование : сборник научных трудов IV Международной научно-практической конференции «Лингвистические и культурологические аспекты современного инженерного образования» памяти кандидата педагогических наук, доцента Н.А. Качалова, Томск, 15-17 ноября 2023 г. — Томск : Изд-во ТПУ, 2023. — С. 141-148. |
| Abstract: | This study explores the performance of commercially available large language models in common language learning tasks. Structure and working principles of neural networks were considered to hypothesize which tasks would perform better. Experiments were conducted to verify the assumptions. Several variants of task adaptations were compared in tests to discover the most resistant to cheating |
| URI: | http://earchive.tpu.ru/handle/11683/77924 |
| Appears in Collections: | Материалы конференций |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| conference_tpu-2023-C85_V2_p141-148.pdf | 758,12 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.