自主监督的节点表示学习通过节点对邻近对齐
IEEE transactions on pattern analysis and machine intelligence
|January 25, 2024
概括
本研究介绍了新的自我监督节点表示学习方法,这些方法将节点和邻里特征对齐. 这些技术改善了图形表示学习,而不需要标记数据,在节点分类任务上实现了强大的性能.
科学领域:
- 图形神经网络 图形神经网络
- 机器学习 机器学习
- 人工智能的人工智能
背景情况:
- 自主监督学习旨在从未标记的图形数据中创建有效的节点表示.
- 从图形结构中学习上下文信息对于信息表示至关重要.
- 现有的方法在有效地捕捉社区环境和避免代表性崩方面面临挑战.
研究的目的:
- 开发简单而有效的自我监督方法,用于节点表示学习.
- 为了将隐藏的节点表示与它们的邻里上下文对齐.
- 为了解决对比学习框架中的记忆开销和表示崩.
主要方法:
- 提出了一种方法来最大限度地提高节点和邻近表示之间的相互信息,作为图表平滑.
- 介绍了对比学习中线下正面样本选择的拓意识正面采样 (TAPS).
- 开发了一个没有负值的解决方案,使用图形信号失相关性 (GSD) 约束来防止崩和过度平滑.
主要成果:
- 节点对邻近的对齐理论上有助于图形光滑.
- 通过考虑结构依赖,TAPS可实现高效的正取样.
- GSD约束有效地打击代表性崩和过度平滑,统一现有方法.
- 基于MLP的编码器和提出的方法显示了对各种数据集的节点分类有希望的结果.
结论:
- 提出的自我监督方法为学习节点表示提供了简单有效的方法.
- 提出的技术,包括TAPS和GSD,提高图表表示学习效率和稳定性.
- 这些方法在不同的图形结构数据尺度上实现了具有竞争力的节点分类性能.
相关概念视频
Associative Learning
370
Associative learning is a fundamental concept in behavioral psychology, wherein a connection is established between two stimuli or events, leading to a learned response. This process is critical in understanding how behaviors are acquired and modified. Conditioning, the mechanism through which associations are formed, can be divided into two main types: classical conditioning and operant conditioning, each elucidating different aspects of associative learning.
Classical conditioning, also known...
Classical conditioning, also known...
370
Node Analysis for AC Circuits
320
Consider an angioplasty system featuring a catheter equipped with a turbine, a critical tool for removing plaque deposits from coronary arteries. This intricate medical device operates using a circuit model reminiscent of a dual-node RLC circuit powered by a current-controlled voltage source.
To unravel the complexities of this system, nodal analysis is employed, a powerful technique founded on Kirchhoff's current law (KCL), which remains valid for phasors. AC circuits can effectively be...
To unravel the complexities of this system, nodal analysis is employed, a powerful technique founded on Kirchhoff's current law (KCL), which remains valid for phasors. AC circuits can effectively be...
320
State Space Representation
208
The frequency-domain technique, commonly used in analyzing and designing feedback control systems, is effective for linear, time-invariant systems. However, it falls short when dealing with nonlinear, time-varying, and multiple-input multiple-output systems. The time-domain or state-space approach addresses these limitations by utilizing state variables to construct simultaneous, first-order differential equations, known as state equations, for an nth-order system.
Consider an RLC circuit, a...
Consider an RLC circuit, a...
208
Nodal Analysis
916
Nodal analysis is a fundamental method in electrical engineering used to simplify the process of circuit analysis. This method revolves around the concept of using node voltages as the primary variables for circuit analysis. The objective is to determine the voltage at each node in a circuit, which can then be used to find other quantities of interest, such as currents through specific components.
Consider, for instance, a simple circuit composed of three nodes and three resistors, as shown in...
Consider, for instance, a simple circuit composed of three nodes and three resistors, as shown in...
916
Neuron Structure
12.9K
Neurons are the main type of cell in the nervous system that generate and transmit electrochemical signals. They primarily communicate with each other using neurotransmitters at specific junctions called synapses. Neurons come in many shapes that often relate to their function, but most share three main structures: an axon and dendrites that extend out from a cell body.
Structure and Function of Neurons
The neuronal cell body—the soma— houses the nucleus and organelles vital to...
Structure and Function of Neurons
The neuronal cell body—the soma— houses the nucleus and organelles vital to...
12.9K
Structural Classification of Joints
3.4K
Joints, also known as articulations, are classified based on their structural characteristics, i.e., based on whether the articulating surfaces of the adjacent bones are directly connected by fibrous connective tissue or cartilage, or whether the articulating surfaces contact each other within a fluid-filled joint cavity. These differences serve to divide the joints of the body into three structural classifications.
A fibrous joint is where the adjacent bones are united by fibrous connective...
A fibrous joint is where the adjacent bones are united by fibrous connective...
3.4K


