Three-way concept lattice from adjunctive positive and negative concepts

International Journal of Approximate Reasoning, Год журнала: 2024, Номер 174, С. 109272 - 109272

Опубликована: Авг. 23, 2024

Язык: Английский

---

Multigranularity Data Analysis With Zentropy Uncertainty Measure for Efficient and Robust Feature Selection

IEEE Transactions on Cybernetics, Год журнала: 2024, Номер 55(2), С. 740 - 752

Опубликована: Дек. 3, 2024

Multigranularity data analysis has recently become an active research topic in the intelligent computing and mining fields. Feature selection via multigranularity is effective tool for characterizing hierarchical enhancing accuracy of results. Although method been widely adopted feature selection, existing studies still present one prevalent disadvantage: mostly focuses on information presented at a single granularity while ignoring structure data, which contrary to nature multigranularity. Hence, this article proposes with zentropy uncertainty measure efficient robust selection. Specifically, consistent degree first introduced obtain optimal combinations establish neighborhood model processing. Then, novel developed by integrating information, namely zentropy-based measure. Considering its among measures, two important measures are further designed applied Extensive experiments demonstrate that proposed can achieve better robustness classification performance than other state-of-the-art methods.

Язык: Английский

---

The relationship among some special concepts from the perspective of formal context restoration

Applied Intelligence, Год журнала: 2025, Номер 55(4)

Опубликована: Янв. 13, 2025

Язык: Английский

---

Parameter efficient face frontalization in image sequences via GAN inversion

IET Image Processing, Год журнала: 2025, Номер 19(1)

Опубликована: Янв. 1, 2025

Abstract Processing facial images with varying poses is a significant challenge. Most existing face frontalization methods rely on heavy architectures that struggle small datasets and produce low‐quality images. Additionally, although video frames provide richer information, these typically use single due to the lack of suitable multi‐image datasets. To address issues, parameter‐efficient framework for high‐quality in both multi‐frame scenarios proposed. First, high‐quality, diverse dataset created tasks. Second, novel single‐image method introduced by combining GAN inversion transfer learning. This approach reduces number trainable parameters over 91% compared while achieving far more photorealistic results than GAN‐based methods. Finally, this extended sequences images, using attention mechanisms merge information from multiple frames. artefacts like eye blinks improves reconstruction quality. Experiments demonstrate outperforms pSp, state‐of‐the‐art method, 0.15 LPIPS improvement 0.10 increase ID similarity. further identity preservation 0.87, showcasing its effectiveness frontal‐view reconstructions.

Язык: Английский

---

Concept cognition over knowledge graphs: A perspective from mining multi-granularity attribute characteristics of concepts

Information Processing & Management, Год журнала: 2025, Номер 62(4), С. 104095 - 104095

Опубликована: Фев. 11, 2025

Язык: Английский

---