高信頼識別のための最適リジェクトの理論および応用研究

高可靠识别最优拒绝的理论与应用研究

• 批准号: 22KJ2398
• 负责人: JI XIAOTONG
• 金额: $ 1.41万
• 依托单位: Kyushu University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for JSPS Fellows
• 财政年份: 2023
• 资助国家: 日本
• 起止时间: 2023-03-08 至 2024-03-31
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-22KJ2398/
• 关键词: PCF; Learning with rejection; Top-rank learning; Writer-independent; Signature verification; rejection; machine learning; CNN; signature verification

My research focuses on improving the reliability of machine learning models, particularly in scenarios where mistakes are unacceptable, such as medical image recognition and signature verification. Since no model can guarantee perfection, it is essential to improve their reliability under the assumption that mistakes may occur. In this report, I discuss my research from two perspectives: rejection operation and top-rank learning.The rejection operation removes samples that significantly impact the recognition performance, such as those with ambiguous confidence scores. These samples may arise from under-learning or be intrinsically unable to be classified into a single class.In contrast to the rejection operation, top-rank learning aims to improve the model's reliability from an “absolute” perspective. This methodology is more suitable for applications requiring high reliability rather than overall performance, such as identifying patients who “absolutely” do not have cancer from patients with a slight chance of having cancer. I propose a novel machine-learning framework to achieve this goal, and this work is under review.I applied the rejection operation and top-rank learning methodologies to a writer-independent signature verification task for my recent research achievement. The proposed framework improved the model’s reliability, and this is the first application of these two machine-learning frameworks to highly reliable signature verification. The results were quantitative and qualitatively robust.

我的研究重点是提高机器学习模型的可靠性，特别是在错误不可接受的情况下，例如医学图像识别和签名验证。由于任何模型都不能保证完美，因此在假设可能出现错误的情况下，提高模型的可靠性至关重要。在这篇报告中，我从拒绝操作和顶级学习两个角度来讨论我的研究。拒绝操作去除显著影响识别性能的样本，例如那些具有模糊置信度分数的样本。这些样本可能来自学习不足或本质上无法归类为单一类别。与拒绝操作相比，top-rank学习旨在从“绝对”的角度提高模型的可靠性。这种方法更适合于需要高可靠性而不是整体性能的应用，例如从患癌症的可能性很小的患者中识别“绝对”没有患癌症的患者。我提出了一个新的机器学习框架来实现这一目标，这项工作正在审查中。在我最近的研究成果中，我将拒绝操作和顶级学习方法应用于一个独立于作者的签名验证任务。提出的框架提高了模型的可靠性，这是这两种机器学习框架首次应用于高可靠的签名验证。结果是定量和定性稳健的。

项目成果

情報学専攻の博士学生が見た 最先端AI（機械学習）の 実装技術

信息学博士生眼中的前沿AI（机器学习）实现技术

• 发表时间: 2022
• 作者: 椋本浩太;辻健;池田達紀;上杉次郎;Mukumoto Kota and Tsuji Takeshi;XIAOTONG JI;XIAOTONG JI;XIAOTONG JI
• 通讯作者: XIAOTONG JI

Revealing Reliable Signatures by Learning Top-Rank Pairs

通过学习顶级对来揭示可靠的签名

• DOI: 10.1007/978-3-031-06555-2_22
• 发表时间: 2022
• 期刊: Proceedings of the 15th IAPR International Workshop on Document Analysis Systems (DAS2022)
• 作者: Ji Xiaotong;Zheng Yan;Suehiro Daiki;Uchida Seiichi
• 通讯作者: Uchida Seiichi

Learning Top-Rank Pairs Discloses Reliable Signatures in Writer-Independent Signature Verification

学习顶级对揭示了独立​​于作者的签名验证中的可靠签名

• 发表时间: 2022
• 作者: 椋本浩太;辻健;池田達紀;上杉次郎;Mukumoto Kota and Tsuji Takeshi;XIAOTONG JI;XIAOTONG JI
• 通讯作者: XIAOTONG JI