全球知名科研期刊《Science》近日宣布，为打击学术造假，该期刊开始引入一款名为Proofig的商用AI软件。这款软件自称能够自动识别论文中的欺诈性图像，从而确保学术成果的公正性与真实性。
《Science》期刊团队表示，他们已经对Proofig软件进行了数月的测试。据称，该AI软件能够准确识别经过裁剪、旋转、更改对比度等处理的图像，并从中圈出可能存在欺诈性的图像。这些图像随后将交由人类编辑进一步检查，以确保论文的质量。
引入AI技术检测图像欺诈，是《Science》期刊为打击学术不端行为所采取的一项重要举措。此举有望提高期刊的学术信誉，维护科研领域的公平竞争环境。
Title: Science journal introduces AI to detect fraudulent images, combat academic fraud
Keywords: Science, AI, image fraud
News content:
In a move to combat academic fraud, the renowned scientific journal Science has recently announced the introduction of a commercial AI software called Proofig. The software claims to be able to automatically identify fraudulent images in papers, ensuring the fairness and authenticity of academic achievements.
The Science team said that they have been testing the Proofig software for several months. The AI tool is said to be able to accurately detect images that have been cropped, rotated, or adjusted in contrast, and flag potential fraudulent images for further review by human editors.
By introducing AI technology to detect image fraud, the journal is taking an important step to combat academic misconduct. This measure is expected to enhance the academic credibility of Science and maintain a fair competitive environment in the research field.
【来源】https://www.ithome.com/0/743/509.htm
Views: 1