Computing Graph Neural Networks: A Survey from Algorithms to Accelerators

Abadal, Sergi; Jain, Akshay; Guirado, Robert; López-Alonso, Jorge; Alarcón, Eduard

doi:10.48550/arxiv.2010.00130

Cited by 5 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…How to effectively compute GNNs in order to realise their full potential will be a key research topic in the coming years. Several hardware accelerators have been developed to cope with GNNs' high density and alternating computing requirements, but there is not a clear proposal applicable to multiple GNN variants [17]. On the software side, current deep learning frameworks including extensions of popular libraries such as TensorFlow and PyTorch have limitations when implementing dynamic computation graphs along with specialized tensor operations [17].…”

Section: B Challenges In Adapting Graph-based Deep Learning Methods F...mentioning

confidence: 99%

“…Several hardware accelerators have been developed to cope with GNNs' high density and alternating computing requirements, but there is not a clear proposal applicable to multiple GNN variants [17]. On the software side, current deep learning frameworks including extensions of popular libraries such as TensorFlow and PyTorch have limitations when implementing dynamic computation graphs along with specialized tensor operations [17]. Thus, there is a need to further develop libraries such as DGL [246] which may handle the sparsity of GNN operations efficiently, as well as complex tensor operations in CUDA with GPU computation acceleration.…”

Section: B Challenges In Adapting Graph-based Deep Learning Methods F...mentioning

confidence: 99%

See 1 more Smart Citation

Graph-Based Deep Learning for Medical Diagnosis and Analysis: Past, Present and Future

Ahmedt-Aristizabal,

Armin,

Denman

et al. 2021

Preprint

View full text Add to dashboard Cite

With the advances of data-driven machine learning research, a wide variety of prediction problems have been tackled. It has become critical to explore how machine learning and specifically deep learning methods can be exploited to analyse healthcare data. A major limitation of existing methods has been the focus on grid-like data; however, the structure of physiological recordings are often irregular and unordered which makes it difficult to conceptualise them as a matrix. As such, graph neural networks have attracted significant attention by exploiting implicit information that resides in a biological system, with interactive nodes connected by edges whose weights can be either temporal associations or anatomical junctions. In this survey, we thoroughly review the different types of graph architectures and their applications in healthcare. We provide an overview of these methods in a systematic manner, organized by their domain of application including functional connectivity, anatomical structure and electrical-based analysis. We also outline the limitations of existing techniques and discuss potential directions for future research.

show abstract

Section: B Challenges In Adapting Graph-based Deep Learning Methods F...mentioning

confidence: 99%

Section: B Challenges In Adapting Graph-based Deep Learning Methods F...mentioning

confidence: 99%

Graph-Based Deep Learning for Medical Diagnosis and Analysis: Past, Present and Future

Ahmedt-Aristizabal,

Armin,

Denman

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…predictions) for either a node, an edge, or the entire graph. The GNN inference process is generally divided into two computation stages, namely, aggregation and combination [5]. First, the data from nodes or edges is loaded from the memory hierarchy to the Processing Elements (PEs).…”

Section: Accelerators Algorithms Supportedmentioning

confidence: 99%

“…Representation of data-sets as graph-based data structures, given the opportunities they provision to understand complex relationships embedded in them, has become increasingly popular [1]- [5]. These graph representations can range from very small (chemistry) to extremely huge (recommendation systems) graphs [6], [7].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Characterizing the Communication Requirements of GNN Accelerators: A Model-Based Approach

Guirado

Jain

Abadal

et al. 2021

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

Self Cite

View full text Add to dashboard Cite

Relational data present in real world graph representations demands for tools capable to study it accurately. In this regard Graph Neural Network (GNN) is a powerful tool, wherein various models for it have also been developed over the past decade. Recently, there has been a significant push towards creating accelerators that speed up the inference and training process of GNNs. These accelerators, however, do not delve into the impact of their dataflows on the overall data movement and, hence, on the communication requirements. In this paper, we formulate analytical models that capture the amount of data movement in the most recent GNN accelerator frameworks. Specifically, the proposed models capture the dataflows and hardware setup of these accelerator designs and expose their scalability characteristics for a set of hardware, GNN model and input graph parameters. Additionally, the proposed approach provides means for the comparative analysis of the vastly different GNN accelerators.

show abstract

A Taxonomy for Classification and Comparison of Dataflows for GNN Accelerators

Garg¹,

Qin²,

Muñoz-Martínez³

et al. 2021

Self Cite

View full text Add to dashboard Cite

Recently, Graph Neural Networks (GNNs) have received a lot of interest because of their success in learning representations from graph structured data. However, GNNs exhibit different compute and memory characteristics compared to traditional Deep Neural Networks (DNNs). Graph convolutions require feature aggregations from neighboring nodes (known as the aggregation phase), which leads to highly irregular data accesses. GNNs also have a very regular compute phase that can be broken down to matrix multiplications (known as the combination phase). All recently proposed GNN accelerators utilize different dataflows and microarchitecture optimizations for these two phases. Different communication strategies between the two phases have been also used. However, as more custom GNN accelerators are proposed, the harder it is to qualitatively classify them and quantitatively contrast them. In this work, we present a taxonomy to describe several diverse dataflows for running GNN inference on accelerators. This provides a structured way to describe and compare the design-space of GNN accelerators.

show abstract

Computing Graph Neural Networks: A Survey from Algorithms to Accelerators

Cited by 5 publications

References 0 publications

Graph-Based Deep Learning for Medical Diagnosis and Analysis: Past, Present and Future

Graph-Based Deep Learning for Medical Diagnosis and Analysis: Past, Present and Future

Characterizing the Communication Requirements of GNN Accelerators: A Model-Based Approach

A Taxonomy for Classification and Comparison of Dataflows for GNN Accelerators

Contact Info

Product

Resources

About