Partial label learning: Taxonomy, analysis and outlook

Tian, Yingjie; Yu, Xiaotong; Fu, Saiji

doi:10.1016/j.neunet.2023.02.019

Cited by 20 publications

(3 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure 6b illustrates this setting, which is known as multi-dimensional partial label classification [105] and deals with the problem where we only know a candidate label set containing the ground-truth class label in each dimension. It can also be regarded as a multi-dimensional extension to partial label learning [106].…”

Section: Multi-label Classificationmentioning

confidence: 99%

Multi-dimensional classification: paradigm, algorithms and beyond

Jia,

Zhang

2024

Vicinagearth

View full text Add to dashboard Cite

Multi-dimensional classification (MDC) aims at learning from objects where each of them is represented by a single instance while associated with multiple class variables. In recent years, this practical learning paradigm has attracted increasing attentions in machine learning community. In this paper, a timely review on this topic is provided with emphasis on representative algorithms. Firstly, the MDC learning framework, commonly used evaluation metrics and publicly available MDC datasets are given. Then, eight state-of-the-art MDC algorithms are scrutinized as the representatives of three categories. After that, several related learning settings are briefly summarized. Finally, this paper is concluded with discussing some open problems to be studied in the future.

show abstract

Section: Multi-label Classificationmentioning

confidence: 99%

Multi-dimensional classification: paradigm, algorithms and beyond

Jia,

Zhang

2024

Vicinagearth

View full text Add to dashboard Cite

show abstract

“…Previous methods (Wang et al 2021;Wu, Wang, and Zhang 2022;Lv et al 2020;Xia et al 2023;Lyu, Wu, and Feng 2022;Zhang et al 2021) for deep PLL achieved remarkable performance which even can be comparable to supervised learning under certain experiment settings. However, existing methods based on deep learning mainly focus on synthetic PLL datasets which are generated from generic datasets (e.g., CIFAR-10/100) (Tian, Yu, and Fu 2023). The categories involved in these synthetic PLL datasets include various objects, such as bird, car, plane and so on.…”

Section: Introductionmentioning

confidence: 99%

Navigating Real-World Partial Label Learning: Unveiling Fine-Grained Images with Attributes

Jiang,

Sun,

Tian

2024

AAAI

View full text Add to dashboard Cite

Partial label learning (PLL), a significant research area, addresses the challenge of annotating each sample with a candidate label set containing the true label when obtaining accurate labels is infeasible. However, existing PLL methods often rely on generic datasets like CIFAR, where annotators can readily differentiate candidate labels and are unlikely to confuse, making it less realistic for real-world partial label applications. In response, our research focuses on a rarely studied problem, PLL on fine-grained images with attributes. And we propose a novel framework called Shared to Learn, Distinct to Disambiguate (SoDisam). Within the candidate label set, the categories may exhibit numerous shared attribute features, posing a challenge in accurately distinguishing them. Rather than perceiving it as an impediment, we capitalize on these shared attributes as definitive sources of supervision. This insight guides us to learn attribute space visual representation to focus on the information from these shared attributes. Moreover, we introduce an attribute attention mechanism tailored to harness the remaining distinct attributes. This mechanism directs the originally holistic feature towards specific regions, capturing corresponding discriminative features. In addition, a dynamic disambiguation module is introduced, continuously adjusting the two aforementioned mechanisms and achieve the final disambiguation process. Extensive experiments demonstrate the effectiveness of our approach on fine-grained partial label datasets. The proposed SoDisam framework not only addresses the challenges associated with fine-grained partial label learning but also provides a more realistic representation of real-world partial label scenarios.

show abstract

“…Using PLL can significantly reduce the time and money costs of labeling large amounts of data while ensuring model accuracy. Currently, PLL can be divided into four types: transformation strategy, theory-oriented strategy, extensions, and disambiguation strategy [28]. The transformation strategy includes binary learning, graph matching, and dictionary learning.…”

Section: Introductionmentioning

confidence: 99%

Blood Cell Attribute Classification Algorithm Based on Partial Label Learning

Feng,

Guo,

Luo

et al. 2024

Electronics

View full text Add to dashboard Cite

Hematological morphology examinations, essential for diagnosing blood disorders, increasingly utilize deep learning. Blood cell classification, determined by combinations of cell attributes, is complicated by the complex relationships and subtle differences among the attributes, resulting in significant time and cost penalties. This study introduces the Partial Label Learning for Blood Cell Classification (P4BC) strategy, a method that trains neural networks using the blood cell attribute labeling data of weak annotations. Using morphological knowledge, we predefined candidate label sets for the blood cell attributes to blend this knowledge with deep learning. This improves the model’s prediction accuracy and interpretability in classifying attributes. This method effectively combines morphological knowledge with deep learning, an approach we refer to as knowledge alignment. It results in an 8.66% increase in attribute recognition accuracy and a 1.09% improvement in matching predictions to the candidate label sets, compared to the original method. These results confirm our method’s ability to grasp the characteristic information of blood cell attributes, enhancing the model interpretability and achieving knowledge alignment between hematological morphology and deep learning. Our algorithm ensures attribute classification accuracy and shows excellent cell category classification, highlighting its wide application potential and practical value in blood cell category classification.

show abstract

Partial label learning: Taxonomy, analysis and outlook

Cited by 20 publications

References 61 publications

Multi-dimensional classification: paradigm, algorithms and beyond

Multi-dimensional classification: paradigm, algorithms and beyond

Navigating Real-World Partial Label Learning: Unveiling Fine-Grained Images with Attributes

Blood Cell Attribute Classification Algorithm Based on Partial Label Learning

Contact Info

Product

Resources

About