中文多模態圖文不符資料集;Chinese Multi-modal Out-of-context Dataset

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/92987

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/92987

Title:	中文多模態圖文不符資料集;Chinese Multi-modal Out-of-context Dataset
Authors:	林斳;Lin, Chyn
Contributors:	資訊工程學系
Keywords:	圖文不符;資料集;Out-of-context;Dataset
Date:	2023-02-02
Issue Date:	2024-09-19 16:36:45 (UTC+8)
Publisher:	國立中央大學
Abstract:	假新聞的驗證一直是普遍存在的問題。近年來隨著社群媒體的發展，訊息傳播更加容易，其中將真實圖片的標題，抽換成誤導訊息的作法，能夠低成本製作具有一定說服力的假新聞，而這種假新聞類型我們稱作圖文不符。在圖文不符方面，過去有人運用文字、圖片和場景變換，從真實新聞製作過英文的圖文不符資料集。但目前仍然沒有中文的圖文不符資料集以及中文模型。於是我們決定創造具有挑戰性且可用的非隨機圖文匹配技術，用於自動生成中文假新聞資料集。並將用於英文圖文不符判斷的模型搬遷至中文，以測試資料集的可用性，並分析了將英文模行搬遷至中文後的性能。;The verification of fake news has always been a common problem. With the development of social media in recent years, it has become easier to disseminate information. The practice of replacing the captions of real pictures with misleading information can produce convincing fake news at a low cost. We call this type of fake news out-of-context. In terms of out-of-context, some people have used text, pictures and scene changes to produce English out-of-context datasets from real news in the past. However, there is still no Chinese out-of-context dataset or Chinese out-ofcontext model. So we decide to create a challenging and usable non-random out-of-context matching technique to automatically generate Chinese out-of-context datasets. We further adapt the model designed for English out-of-context determination to Chinese to test the usability of the dataset, and analyzed the performance of the adapted English model to Chinese.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	7	View/Open

社群 sharing

Loading...