ویراساینس.
مقاله‌ی دسترسی آزاد

Multi-Augmentation Contrastive Learning as Multi-Objective Optimization for Graph Neural Networks

['Xu Li', 'Yongsheng Chen', '']

چکیده

Abstract Recently self-supervised learning is gaining popularity for Graph Neural Networks (GNN) by leveraging unlabeled data. Augmentation plays a key role in self-supervision. While there common set of image augmentation methods that preserve labels general, graph do not guarantee consistent semantics and are usually domain dependent. Existing GNN models often handpick small techniques limit the performance model. In this paper, we propose to wide range tasks, rely on Pareto optimality select balance among these possibly conflicting augmented versions, called P areto G raph C ontrastive L earning ( PGCL ) framework. We show while random selection same leads slow convergence or even divergence, converges much faster with lower error rate. Extensive experiments multiple datasets different domains scales demonstrate superior comparable PGCL.

اسناد مشابه

  • Molecular contrastive learning of representations via graph neural networks

    مقاله‌ی دسترسی آزاد۲۰۲۲Nature Machine Intelligence['Yuyang Wang', 'Jianren Wang', 'Zhonglin Cao', 'Amir Barati Farimani', '']

    Molecular Machine Learning (ML) bears promise for efficient molecule property prediction and drug discovery. However, labeled data can be expensive time-consuming to acquire. Due the limited data, it is a great challenge supervised-learning ML models generalize giant chemical spa

  • ALEC: An Adaptive Learning Framework for Optimizing Artificial Neural Networks

    سند علمی۲۰۰۱International Conference on Computational ScienceAjith Abraham, Baikunth Nath

    In this paper we present ALEC (Adaptive Learning by Evolutionary Computation), an automatic computational framework for optimizing neural networks wherein the neural network topology (architecture, activation function, weights) and learning algorithms are adapted according to the

  • Multi-criterion Pareto based particle swarm optimized polynomial neural network for classification: A review and state-of-the-art

    سند علمی۲۰۰۹Computer Science ReviewSatchidananda Dehuri, Sung-Bae Cho

    In this paper, we proposed a multi-objective Pareto based particle swarm optimization (MOPPSO) to minimize the architectural complexity and maximize the classification accuracy of a polynomial neural network (PNN). To support this, we provide an extensive review of the literature

  • Graph Transformer: Learning Better Representations for Graph Neural Networks

    مقاله‌ی دسترسی آزاد۲۰۲۱Lecture Notes in Computer Science['Boyuan Wang', 'Lixin Cui', 'Lu Bai', 'Edwin R. Hancock', '']

    Graph classifications are significant tasks for many real-world applications. Recently, Neural Networks (GNNs) have achieved excellent performance on graph classification tasks. However, most state-of-the-art GNNs face the challenge of over-smoothing problem and cannot learn late

  • Generalizable Machine Learning in Neuroscience Using Graph Neural Networks

    مقاله‌ی دسترسی آزاد۲۰۲۱Frontiers in artificial intelligence['Paul J. Wang', 'Sandalika Sapra', 'Vivek Kurien George', 'Gabriel A. Silva', '']

    Although a number of studies have explored deep learning in neuroscience, the application these algorithms to neural systems on microscopic scale, i.e. parameters relevant lower scales organization, remains relatively novel. Motivated by advances whole-brain imaging, we examined

  • Uncertainty estimation in deep learning‐based property models: Graph neural networks applied to the critical properties

    مقاله‌ی دسترسی آزاد۲۰۲۲Aiche Journal['Adem R.N. Aouichaoui', 'Seyed Soheil Mansouri', 'Jens Abildskov', 'Gürkan Sin', '']

    Deep learning and graph-based models have gained popularity in various life science applications such as property modeling, achieving state-of-the-art performance. However, the quantification of prediction uncertainty these is less studied not applied low dataset size regime, whi