Searcharxiv⌕ Search

arXiv subjects

Kijung Shin

Publications and source records attributed to Kijung Shin.

At least 73 records · Page 4Linked to original sources
arXiv2023

Robust Graph Clustering via Meta Weighting for Noisy Graphs

How can we find meaningful clusters in a graph robustly against noise edges? Graph clustering (i.e., dividing nodes into groups of similar ones) is a fundamental problem in graph analysis with applications in various fields. Recent studies have demonstrated that graph neural network (GNN) based approaches yield promising results for graph clustering. However, we observe that their performance degenerates significantly on graphs with noise edges, which are prevalent in practice. In this work, we propose MetaGC for robust GNN-based graph clustering. MetaGC employs a decomposable clustering loss function, which can be rephrased as a sum of losses over node pairs. We add a learnable weight to each node pair, and MetaGC adaptively adjusts the weights of node pairs using meta-weighting so that the weights of meaningful node pairs increase and the weights of less-meaningful ones (e.g., noise edges) decrease. We show empirically that MetaGC learns weights as intended and consequently outperforms the state-of-the-art GNN-based competitors, even when they are equipped with separate denoising schemes, on five real-world graphs under varying levels of noise. Our code and datasets are available at https://github.com/HyeonsooJo/MetaGC.

cs.LG↗
arXiv2023

How Transitive Are Real-World Group Interactions? -- Measurement and Reproduction

Many real-world interactions (e.g., researcher collaborations and email communication) occur among multiple entities. These group interactions are naturally modeled as hypergraphs. In graphs, transitivity is helpful to understand the connections between node pairs sharing a neighbor, and it has extensive applications in various domains. Hypergraphs, an extension of graphs, are designed to represent group relations. However, to the best of our knowledge, there has been no examination regarding the transitivity of real-world group interactions. In this work, we investigate the transitivity of group interactions in real-world hypergraphs. We first suggest intuitive axioms as necessary characteristics of hypergraph transitivity measures. Then, we propose a principled hypergraph transitivity measure HyperTrans, which satisfies all the proposed axioms, with a fast computation algorithm Fast-HyperTrans. After that, we analyze the transitivity patterns in real-world hypergraphs distinguished from those in random hypergraphs. Lastly, we propose a scalable hypergraph generator THera. It reproduces the observed transitivity patterns by leveraging community structures, which are pervasive in real-world hypergraphs. Our code and datasets are available at https://github.com/kswoo97/hypertrans.

cs.SI↗
arXiv2023

Hypergraph Motifs and Their Extensions Beyond Binary

Hypergraphs naturally represent group interactions, which are omnipresent in many domains: collaborations of researchers, co-purchases of items, and joint interactions of proteins, to name a few. In this work, we propose tools for answering the following questions: (Q1) what are the structural design principles of real-world hypergraphs? (Q2) how can we compare local structures of hypergraphs of different sizes? (Q3) how can we identify domains from which hypergraphs are? We first define hypergraph motifs (h-motifs), which describe the overlapping patterns of three connected hyperedges. Then, we define the significance of each h-motif in a hypergraph as its occurrences relative to those in properly randomized hypergraphs. Lastly, we define the characteristic profile (CP) as the vector of the normalized significance of every h-motif. Regarding Q1, we find that h-motifs' occurrences in 11 real-world hypergraphs from 5 domains are clearly distinguished from those of randomized hypergraphs. Then, we demonstrate that CPs capture local structural patterns unique to each domain, and thus comparing CPs of hypergraphs addresses Q2 and Q3. The concept of CP is extended to represent the connectivity pattern of each node or hyperedge as a vector, which proves useful in node classification and hyperedge prediction. Our algorithmic contribution is to propose MoCHy, a family of parallel algorithms for counting h-motifs' occurrences in a hypergraph. We theoretically analyze their speed and accuracy and show empirically that the advanced approximate version MoCHy-A+ is more accurate and faster than the basic approximate and exact versions, respectively. Furthermore, we explore ternary hypergraph motifs that extends h-motifs by taking into account not only the presence but also the cardinality of intersections among hyperedges. This extension proves beneficial for all previously mentioned applications.

cs.SI↗
arXiv2023

TensorCodec: Compact Lossy Compression of Tensors without Strong Data Assumptions

Many real-world datasets are represented as tensors, i.e., multi-dimensional arrays of numerical values. Storing them without compression often requires substantial space, which grows exponentially with the order. While many tensor compression algorithms are available, many of them rely on strong data assumptions regarding its order, sparsity, rank, and smoothness. In this work, we propose TENSORCODEC, a lossy compression algorithm for general tensors that do not necessarily adhere to strong input data assumptions. TENSORCODEC incorporates three key ideas. The first idea is Neural Tensor-Train Decomposition (NTTD) where we integrate a recurrent neural network into Tensor-Train Decomposition to enhance its expressive power and alleviate the limitations imposed by the low-rank assumption. Another idea is to fold the input tensor into a higher-order tensor to reduce the space required by NTTD. Finally, the mode indices of the input tensor are reordered to reveal patterns that can be exploited by NTTD for improved approximation. Our analysis and experiments on 8 real-world datasets demonstrate that TENSORCODEC is (a) Concise: it gives up to 7.38x more compact compression than the best competitor with similar reconstruction error, (b) Accurate: given the same budget for compressed size, it yields up to 3.33x more accurate reconstruction than the best competitor, (c) Scalable: its empirical compression time is linear in the number of tensor entries, and it reconstructs each entry in logarithmic time. Our code and datasets are available at https://github.com/kbrother/TensorCodec.

cs.LG↗
arXiv2023

Reciprocity in Directed Hypergraphs: Measures, Findings, and Generators

Group interactions are prevalent in a variety of areas. Many of them, including email exchanges, chemical reactions, and bitcoin transactions, are directional, and thus they are naturally modeled as directed hypergraphs, where each hyperarc consists of the set of source nodes and the set of destination nodes. For directed graphs, which are a special case of directed hypergraphs, reciprocity has played a key role as a fundamental graph statistic in revealing organizing principles of graphs and in solving graph learning tasks. For general directed hypergraphs, however, even no systematic measure of reciprocity has been developed. In this work, we investigate the reciprocity of 11 real-world hypergraphs. To this end, we first introduce eight axioms that any reasonable measure of reciprocity should satisfy. Second, we propose HyperRec, a family of principled measures of hypergraph reciprocity that satisfies all the axioms. Third, we develop Ferret, a fast and exact algorithm for computing the measure, whose search space is up to 10^{147}x smaller than that of naive computation. Fourth, using them, we examine 11 real-world hypergraphs and discover patterns that distinguish them from random hypergraphs. Lastly, we propose ReDi, an intuitive generative model for directed hypergraphs exhibiting the patterns.

cs.SI↗
arXiv2023

On Improving the Cohesiveness of Graphs by Merging Nodes: Formulation, Analysis, and Algorithms

Graphs are a powerful mathematical model, and they are used to represent real-world structures in various fields. In many applications, real-world structures with high connectivity and robustness are preferable. For enhancing the connectivity and robustness of graphs, two operations, adding edges and anchoring nodes, have been extensively studied. However, merging nodes, which is a realistic operation in many scenarios (e.g., bus station reorganization, multiple team formation), has been overlooked. In this work, we study the problem of improving graph cohesiveness by merging nodes. First, we formulate the problem mathematically using the size of the $k$-truss, for a given $k$, as the objective. Then, we prove the NP-hardness and non-modularity of the problem. After that, we develop BATMAN, a fast and effective algorithm for choosing sets of nodes to be merged, based on our theoretical findings and empirical observations. Lastly, we demonstrate the superiority of BATMAN over several baselines, in terms of speed and effectiveness, through extensive experiments on fourteen real-world graphs.

cs.SI↗
arXiv2023

Classification of Edge-dependent Labels of Nodes in Hypergraphs

A hypergraph is a data structure composed of nodes and hyperedges, where each hyperedge is an any-sized subset of nodes. Due to the flexibility in hyperedge size, hypergraphs represent group interactions (e.g., co-authorship by more than two authors) more naturally and accurately than ordinary graphs. Interestingly, many real-world systems modeled as hypergraphs contain edge-dependent node labels, i.e., node labels that vary depending on hyperedges. For example, on co-authorship datasets, the same author (i.e., a node) can be the primary author in a paper (i.e., a hyperedge) but the corresponding author in another paper (i.e., another hyperedge). In this work, we introduce a classification of edge-dependent node labels as a new problem. This problem can be used as a benchmark task for hypergraph neural networks, which recently have attracted great attention, and also the usefulness of edge-dependent node labels has been verified in various applications. To tackle this problem, we propose WHATsNet, a novel hypergraph neural network that represents the same node differently depending on the hyperedges it participates in by reflecting its varying importance in the hyperedges. To this end, WHATsNet models the relations between nodes within each hyperedge, using their relative centrality as positional encodings. In our experiments, we demonstrate that WHATsNet significantly and consistently outperforms ten competitors on six real-world hypergraphs, and we also show successful applications of WHATsNet to (a) ranking aggregation, (b) node clustering, and (c) product return prediction.

cs.SI↗
arXiv2023

Towards Deep Attention in Graph Neural Networks: Problems and Remedies

Graph neural networks (GNNs) learn the representation of graph-structured data, and their expressiveness can be further enhanced by inferring node relations for propagation. Attention-based GNNs infer neighbor importance to manipulate the weight of its propagation. Despite their popularity, the discussion on deep graph attention and its unique challenges has been limited. In this work, we investigate some problematic phenomena related to deep graph attention, including vulnerability to over-smoothed features and smooth cumulative attention. Through theoretical and empirical analyses, we show that various attention-based GNNs suffer from these problems. Motivated by our findings, we propose AEROGNN, a novel GNN architecture designed for deep graph attention. AERO-GNN provably mitigates the proposed problems of deep graph attention, which is further empirically demonstrated with (a) its adaptive and less smooth attention functions and (b) higher performance at deep layers (up to 64). On 9 out of 12 node classification benchmarks, AERO-GNN outperforms the baseline GNNs, highlighting the advantages of deep graph attention. Our code is available at https://github.com/syleeheal/AERO-GNN.

cs.LG↗
arXiv2023

Interplay between Topology and Edge Weights in Real-World Graphs: Concepts, Patterns, and an Algorithm

What are the relations between the edge weights and the topology in real-world graphs? Given only the topology of a graph, how can we assign realistic weights to its edges based on the relations? Several trials have been done for edge-weight prediction where some unknown edge weights are predicted with most edge weights known. There are also existing works on generating both topology and edge weights of weighted graphs. Differently, we are interested in generating edge weights that are realistic in a macroscopic scope, merely from the topology, which is unexplored and challenging. To this end, we explore and exploit the patterns involving edge weights and topology in real-world graphs. Specifically, we divide each graph into layers where each layer consists of the edges with weights at least a threshold. We observe consistent and surprising patterns appearing in multiple layers: the similarity between being adjacent and having high weights, and the nearly-linear growth of the fraction of edges having high weights with the number of common neighbors. We also observe a power-law pattern that connects the layers. Based on the observations, we propose PEAR, an algorithm assigning realistic edge weights to a given topology. The algorithm relies on only two parameters, preserves all the observed patterns, and produces more realistic weights than the baseline methods with more parameters.

cs.SI↗
arXiv2023

Hypercore Decomposition for Non-Fragile Hyperedges: Concepts, Algorithms, Observations, and Applications

Hypergraphs are a powerful abstraction for modeling high-order relations, which are ubiquitous in many fields. A hypergraph consists of nodes and hyperedges (i.e., subsets of nodes); and there have been a number of attempts to extend the notion of $k$-cores, which proved useful with numerous applications for pairwise graphs, to hypergraphs. However, the previous extensions are based on an unrealistic assumption that hyperedges are fragile, i.e., a high-order relation becomes obsolete as soon as a single member leaves it. In this work, we propose a new substructure model, called ($k$, $t$)-hypercore, based on the assumption that high-order relations remain as long as at least $t$ fraction of the members remain. Specifically, it is defined as the maximal subhypergraph where (1) every node is contained in at least $k$ hyperedges in it and (2) at least $t$ fraction of the nodes remain in every hyperedge. We first prove that, given $t$ (or $k$), finding the ($k$, $t$)-hypercore for every possible $k$ (or $t$) can be computed in time linear w.r.t the sum of the sizes of hyperedges. Then, we demonstrate that real-world hypergraphs from the same domain share similar ($k$, $t$)-hypercore structures, which capture different perspectives depending on $t$. Lastly, we show the successful applications of our model in identifying influential nodes, dense substructures, and vulnerability in hypergraphs.

cs.SI↗
arXiv2023

Characterization of Simplicial Complexes by Counting Simplets Beyond Four Nodes

Simplicial complexes are higher-order combinatorial structures which have been used to represent real-world complex systems. In this paper, we concentrate on the local patterns in simplicial complexes called simplets, a generalization of graphlets. We formulate the problem of counting simplets of a given size in a given simplicial complex. For this problem, we extend a sampling algorithm based on color coding from graphs to simplicial complexes, with essential technical novelty. We theoretically analyze our proposed algorithm named SC3, showing its correctness, unbiasedness, convergence, and time/space complexity. Through the extensive experiments on sixteen real-world datasets, we show the superiority of SC3 in terms of accuracy, speed, and scalability, compared to the baseline methods. Finally, we use the counts given by SC3 for simplicial complex analysis, especially for characterization, which is further used for simplicial complex clustering, where SC3 shows a strong ability of characterization with domain-based similarity.

cs.SI↗
arXiv2023

NeuKron: Constant-Size Lossy Compression of Sparse Reorderable Matrices and Tensors

Many real-world data are naturally represented as a sparse reorderable matrix, whose rows and columns can be arbitrarily ordered (e.g., the adjacency matrix of a bipartite graph). Storing a sparse matrix in conventional ways requires an amount of space linear in the number of non-zeros, and lossy compression of sparse matrices (e.g., Truncated SVD) typically requires an amount of space linear in the number of rows and columns. In this work, we propose NeuKron for compressing a sparse reorderable matrix into a constant-size space. NeuKron generalizes Kronecker products using a recurrent neural network with a constant number of parameters. NeuKron updates the parameters so that a given matrix is approximated by the product and reorders the rows and columns of the matrix to facilitate the approximation. The updates take time linear in the number of non-zeros in the input matrix, and the approximation of each entry can be retrieved in logarithmic time. We also extend NeuKron to compress sparse reorderable tensors (e.g. multi-layer graphs), which generalize matrices. Through experiments on ten real-world datasets, we show that NeuKron is (a) Compact: requiring up to five orders of magnitude less space than its best competitor with similar approximation errors, (b) Accurate: giving up to 10x smaller approximation error than its best competitors with similar size outputs, and (c) Scalable: successfully compressing a matrix with over 230 million non-zero entries.

cs.LG↗
arXiv2023

I'm Me, We're Us, and I'm Us: Tri-directional Contrastive Learning on Hypergraphs

Although machine learning on hypergraphs has attracted considerable attention, most of the works have focused on (semi-)supervised learning, which may cause heavy labeling costs and poor generalization. Recently, contrastive learning has emerged as a successful unsupervised representation learning method. Despite the prosperous development of contrastive learning in other domains, contrastive learning on hypergraphs remains little explored. In this paper, we propose TriCL (Tri-directional Contrastive Learning), a general framework for contrastive learning on hypergraphs. Its main idea is tri-directional contrast, and specifically, it aims to maximize in two augmented views the agreement (a) between the same node, (b) between the same group of nodes, and (c) between each group and its members. Together with simple but surprisingly effective data augmentation and negative sampling schemes, these three forms of contrast enable TriCL to capture both microscopic and mesoscopic structural information in node embeddings. Our extensive experiments using 13 baseline approaches, five datasets, and two tasks demonstrate the effectiveness of TriCL, and most noticeably, TriCL consistently outperforms not just unsupervised competitors but also (semi-)supervised competitors mostly by significant margins for node classification. The code and datasets are available at https://github.com/wooner49/TriCL.

cs.LG↗
arXiv2023

Graphlets over Time: A New Lens for Temporal Network Analysis

Graphs are widely used for modeling various types of interactions, such as email communications and online discussions. Many of such real-world graphs are temporal, and specifically, they grow over time with new nodes and edges. Counting the instances of each graphlet (i.e., an induced subgraph isomorphism class) has been successful in characterizing local structures of graphs, with many applications. While graphlets have been extended for temporal graphs, the extensions are designed for examining temporally-local subgraphs composed of edges with close arrival times, instead of long-term changes in local structures. In this paper, as a new lens for temporal graph analysis, we study the evolution of distributions of graphlet instances over time in real-world graphs at three different levels (graphs, nodes, and edges). At the graph level, we first discover that the evolution patterns are significantly different from those in random graphs. Then, we suggest a graphlet transition graph for measuring the similarity of the evolution patterns of graphs, and we find out a surprising similarity between the graphs from the same domain. At the node and edge levels, we demonstrate that the local structures around nodes and edges in their early stage provide a strong signal regarding their future importance. In particular, we significantly improve the predictability of the future importance of nodes and edges using the counts of the roles (a.k.a., orbits) that they take within graphlets.

cs.SI↗
arXiv2022

Region-Conditioned Orthogonal 3D U-Net for Weather4Cast Competition

The Weather4Cast competition (hosted by NeurIPS 2022) required competitors to predict super-resolution rain movies in various regions of Europe when low-resolution satellite contexts covering wider regions are given. In this paper, we show that a general baseline 3D U-Net can be significantly improved with region-conditioned layers as well as orthogonality regularizations on 1x1x1 convolutional layers. Additionally, we facilitate the generalization with a bag of training strategies: mixup data augmentation, self-distillation, and feature-wise linear modulation (FiLM). Presented modifications outperform the baseline algorithms (3D U-Net) by up to 19.54% with less than 1% additional parameters, which won the 4th place in the core test leaderboard.

cs.CV↗
arXiv2022

Deep-Learning-Based Precipitation Nowcasting with Ground Weather Station Data and Radar Data

Recently, many deep-learning techniques have been applied to various weather-related prediction tasks, including precipitation nowcasting (i.e., predicting precipitation levels and locations in the near future). Most existing deep-learning-based approaches for precipitation nowcasting, however, consider only radar and/or satellite images as inputs, and meteorological observations collected from ground weather stations, which are sparsely located, are relatively unexplored. In this paper, we propose ASOC, a novel attentive method for effectively exploiting ground-based meteorological observations from multiple weather stations. ASOC is designed to capture temporal dynamics of the observations and also contextual relationships between them. ASOC is easily combined with existing image-based precipitation nowcasting models without changing their architectures. We show that such a combination improves the average critical success index (CSI) of predicting heavy (at least 10 mm/hr) and light (at least 1 mm/hr) rainfall events at 1-6 hr lead times by 5.7%, compared to the original image-based model, using the radar images and ground-based observations around South Korea collected from 2014 to 2020.

physics.ao-ph↗
arXiv2022

HashNWalk: Hash and Random Walk Based Anomaly Detection in Hyperedge Streams

Sequences of group interactions, such as emails, online discussions, and co-authorships, are ubiquitous; and they are naturally represented as a stream of hyperedges. Despite their broad potential applications, anomaly detection in hypergraphs (i.e., sets of hyperedges) has received surprisingly little attention, compared to that in graphs. While it is tempting to reduce hypergraphs to graphs and apply existing graph-based methods, according to our experiments, taking higher-order structures of hypergraphs into consideration is worthwhile. We propose HashNWalk, an incremental algorithm that detects anomalies in a stream of hyperedges. It maintains and updates a constant-size summary of the structural and temporal information about the stream. Using the summary, which is the form of a proximity matrix, HashNWalk measures the anomalousness of each new hyperedge as it appears. HashNWalk is (a) Fast: it processes each hyperedge in near real-time and billions of hyperedges within a few hours, (b) Space Efficient: the size of the maintained summary is a predefined constant, (c) Effective: it successfully detects anomalous hyperedges in real-world hypergraphs.

cs.SI↗
arXiv2022

Set2Box: Similarity Preserving Representation Learning of Sets

Sets have been used for modeling various types of objects (e.g., a document as the set of keywords in it and a customer as the set of the items that she has purchased). Measuring similarity (e.g., Jaccard Index) between sets has been a key building block of a wide range of applications, including, plagiarism detection, recommendation, and graph compression. However, as sets have grown in numbers and sizes, the computational cost and storage required for set similarity computation have become substantial, and this has led to the development of hashing and sketching based solutions. In this work, we propose Set2Box, a learning-based approach for compressed representations of sets from which various similarity measures can be estimated accurately in constant time. The key idea is to represent sets as boxes to precisely capture overlaps of sets. Additionally, based on the proposed box quantization scheme, we design Set2Box+, which yields more concise but more accurate box representations of sets. Through extensive experiments on 8 real-world datasets, we show that, compared to baseline approaches, Set2Box+ is (a) Accurate: achieving up to 40.8X smaller estimation error while requiring 60% fewer bits to encode sets, (b) Concise: yielding up to 96.8X more concise representations with similar estimation error, and (c) Versatile: enabling the estimation of four set-similarity measures from a single representation of each set.

cs.SI↗