Predictive Adversarial Learning from Positive and Unlabeled Data

Hu, Wenpeng; Le, Ran; Liu, Bing; Ji, Feng; Ma, Jinwen; Zhao, Dongyan; Yan, Rui

doi:10.1609/aaai.v35i9.16953

Cited by 22 publications

(16 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An intrinsic challenge for this class of problems is that the number of true positives, i.e. the prior class distribution [86][87][88][89][90][91] , is unknown and most classifiers require labels for training. Motivated by the robustness and the performance of ensemble approaches such as bagging in PU learning 39,86,87 , we develop a statistical approach to separate candidate genes from non-candidate genes using an ensemble approach 87,88,92 which eliminates the need to predefine 39 or estimate 88 a prior class distribution or to choose an arbitrary cut-off 40,42 on predicted rank distributions.…”

Section: The Ensemble Approachmentioning

confidence: 99%

Speos: An ensemble graph representation learning framework to predict core genes for complex diseases

Ratajczak

Joblin

Hildebrandt

et al. 2023

Preprint

View full text Add to dashboard Cite

Understanding phenotype-to-genotype relationships is a grand challenge of 21st century biology with translational implications. The recently proposed 'omnigenic' model postulates that effects of genetic variation on traits are mediated by core-genes and -proteins whose activities mechanistically influence the phenotype, whereas peripheral genes encode a regulatory network that indirectly affects phenotypes via core gene products. We have developed a positive-unlabeled graph representation-learning ensemble-approach to predict core genes for diverse diseases using Mendelian disorder genes for training. Employing mouse knockout phenotypes for external validation, we demonstrate that our most confident predictions validate at rates on par with the Mendelian disorder genes, and all candidates exhibit core-gene properties like transcriptional deregulation in diseases and loss-of-function intolerance. Predicted candidates are enriched for drug targets and druggable proteins and, in contrast to Mendelian disorder genes, also for druggable but yet untargeted gene products. Model interpretation suggests key molecular mechanisms and physical interactions for core gene predictions. Our results demonstrate the potential of graph representation learning and pave the way for studying core gene properties and future drug development.

show abstract

Section: The Ensemble Approachmentioning

confidence: 99%

Speos: An ensemble graph representation learning framework to predict core genes for complex diseases

Ratajczak

Joblin

Hildebrandt

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…For examples, GenPU in (Hou et al 2018) proposed to train an array of generators and discriminators to distinguish between the positive and negative instances in the unlabelled datasets. PAN (Hu et al 2021) proposed to train the PU-learning classifier under the GAN framework by viewing instances selected by the classifier as the generated instances. However, these works generally focus on how to obtain a better classifier/discriminator rather than how to generate high-quality desired instances.…”

Section: Related Workmentioning

confidence: 99%

Leveraging Contaminated Datasets to Learn Clean-Data Distribution with Purified Generative Adversarial Networks

Tian¹,

Su²,

Yu³

2023

Preprint

View full text Add to dashboard Cite

Generative adversarial networks (GANs) are known for their strong abilities on capturing the underlying distribution of training instances. Since the seminal work of GAN, many variants of GAN have been proposed. However, existing GANs are almost established on the assumption that the training dataset is clean. But in many real-world applications, this may not hold, that is, the training dataset may be contaminated by a proportion of undesired instances. When training on such datasets, existing GANs will learn a mixture distribution of desired and contaminated instances, rather than the desired distribution of desired data only (target distribution). To learn the target distribution from contaminated datasets, two purified generative adversarial networks (PuriGAN) are developed, in which the discriminators are augmented with the capability to distinguish between target and contaminated instances by leveraging an extra dataset solely composed of contamination instances. We prove that under some mild conditions, the proposed PuriGANs are guaranteed to converge to the distribution of desired instances. Experimental results on several datasets demonstrate that the proposed PuriGANs are able to generate much better images from the desired distribution than comparable baselines when trained on contaminated datasets. In addition, we also demonstrate the usefulness of PuriGAN on downstream applications by applying it to the tasks of semi-supervised anomaly detection on contaminated datasets and PU-learning. Experimental results show that PuriGAN is able to deliver the best performance over comparable baselines on both tasks 1 .

show abstract

“…Predictive adversarial learning were introduced by [20,42]. They used generator to produce data to fool a discriminator which can determine the generated data is positive or not.…”

Section: Related Work 21 Positive-unlabeled (Pu) Learningmentioning

confidence: 99%

“…The generated data is used to train binary classifier. PAN [20] proposes a new objective based on KL-distance, and optimize the architecture of GAN. In this work, we proposed a PUtree algorithm which splits PU instances into hierarchical communities, and then a PU path fusion network is deployed to aggregate different level community information, which can be vital for various PU tasks, especially for chronic disease prediction.…”

Section: Related Work 21 Positive-unlabeled (Pu) Learningmentioning

confidence: 99%

Community-Based Hierarchical Positive-Unlabeled (PU) Model Fusion for Chronic Disease Prediction

Wu,

Li,

Zhang

et al. 2023

Proceedings of the 32nd ACM International Conference on Information and Knowledge Management

View full text Add to dashboard Cite

Positive-Unlabeled (PU) Learning is a challenge presented by binary classification problems where there is an abundance of unlabeled data along with a small number of positive data instances, which can be used to address chronic disease screening problem. State-ofthe-art PU learning methods have resulted in the development of various risk estimators, yet they neglect the differences among distinct populations. To address this issue, we present a novel Positive-Unlabeled Learning Tree (PUtree) algorithm. PUtree is designed to take into account communities such as different age or income brackets, in tasks of chronic disease prediction. We propose a novel approach for binary decision-making, which hierarchically builds community-based PU models and then aggregates their deliverables. Our method can explicate each PU model on the tree for the optimized non-leaf PU node splitting. Furthermore, a mask-recovery data augmentation strategy enables sufficient training of the model in individual communities. Additionally, the proposed approach includes an adversarial PU risk estimator to capture hierarchical PU-relationships, and a model fusion network that integrates data from each tree path, resulting in robust binary classification results. We demonstrate the superior performance of PUtree as well as its variants on two benchmarks and a new diabetes-prediction dataset. CCS CONCEPTS• Computing methodologies → Semi-supervised learning settings; • Applied computing → Consumer health; • Information systems → Combination, fusion and federated search.

show abstract

Predictive Adversarial Learning from Positive and Unlabeled Data

Cited by 22 publications

References 29 publications

Speos: An ensemble graph representation learning framework to predict core genes for complex diseases

Speos: An ensemble graph representation learning framework to predict core genes for complex diseases

Leveraging Contaminated Datasets to Learn Clean-Data Distribution with Purified Generative Adversarial Networks

Community-Based Hierarchical Positive-Unlabeled (PU) Model Fusion for Chronic Disease Prediction

Contact Info

Product

Resources

About