Next-Generation Big Graph Learning and Analytics (Next-Gen-BiGLA)

Introduction to the Workshop

In today's rapidly evolving digital landscape, the significance of graph-based data structures and algorithms cannot be overstated. Graphs, representing complex networks of data from social media interactions to biological networks and beyond, are fundamental in extracting valuable insights that drive major technological and social advancements. However, as the size and complexity of these graphs continue to expand exponentially, existing graph processing techniques often fall short, unable to effectively support learning and analytic workloads simultaneously or adapt to dynamic changes in real time. Moreover, in an era where data is ubiquitous, ensuring the interoperability, privacy, security, and efficiency of graph-based systems is more critical than ever.

Recognizing the urgent need to address these challenges, we propose the "Next-Generation Big Graph Learning and Analytics (Next-Gen-BiGLA)" workshop. This initiative is dedicated to pioneering advanced solutions in the realm of graph learning and analytics, focusing on scalability, dynamism, and innovative algorithmic approaches. Our workshop will delve into a broad array of cutting-edge topics, including scalable algorithms, graph neural networks, and quantum graph computing, among others.

By convening leading experts and innovators from diverse fields such as bioinformatics, cybersecurity, and social networking, Next-Gen-BiGLA aims to foster a collaborative environment where interdisciplinary approaches and ideas can merge to propel the frontier of graph computing forward. This workshop is not just about theoretical enhancements; it is geared towards realizing practical, scalable solutions that can have a broad socio-economic impact.

Through engaging discussions, collaborative research efforts, and benchmarking the next wave of graph technologies, we envision Next-Gen-BiGLA as a cornerstone event that will shape the future of graph computing. It is an opportunity for the community to share insights, challenge existing paradigms, and collectively advance toward next-generation technologies that harness the full potential of big graph data analytics.

Research Topics Included in the Workshop

• The theoretical foundation for graph management and learning
• Scalable algorithms for graph learning and analytics
• Innovative algorithms and systems for streaming/dynamic graphs
• Graph neural networks and deep learning on graphs
• Bridging foundational models with graph computing
• Temporal graph learning and analytics
• Novel graph applications at large scale
• Multi-model data management that involves graph data
• Quantum algorithms for graph computing and learning
• Applications of scalable graph analytics in domains such as social networks, bioinformatics, and cybersecurity
• Benchmarking and evaluation frameworks for scalable graph processing
• Interoperability and standards for graph data and analytics
• Privacy and security issues in graph learning and analytics
• New hardware/platforms for graph computing
• Energy-saving algorithms/systems for scalable graph computing

Important Dates

Agenda

Next-Generation Big Graph Learning and Analytics (Next-Gen-BiGLA)

Location: CAPITOL A
Time: 14:00-17:00

Invited Talks Abstracts

Dr. Chao Zhang: Incremental Sliding Window Connectivity over Streaming Graphs.

Abstract: We study index-based processing for connectivity queries within sliding windows on streaming graphs. These queries, which determine whether two vertices belong to the same connected component, are fundamental operations in real-time graph data processing and demand high throughput and low latency. While indexing methods that leverage data structures for fully dynamic connectivity can facilitate efficient query processing, they encounter significant challenges with deleting expired edges from the window during window updates. We introduce a novel indexing approach that eliminates the need for physically performing edge deletions. This is achieved through a unique bidirectional incremental computation framework, referred to as the BIC model. The BIC model implements two distinct incremental computations to compute connected components within the window, operating along and against the timeline, respectively. These computations are then merged to efficiently compute queries in the window. We propose techniques for optimized index storage, incremental index updates, and efficient query processing to improve BIC effectiveness. Empirically, BIC achieves a 14× increase in throughput and a reduction in P95 latency by up to 3900× when compared to state-of-the-art indexes.

Dr. Yutong Ye: Learning based Graph Query Optimization

Abstract: In recent years, learning-based graph query optimization has demonstrated great potential in managing large-scale graph data, offering innovative solutions for efficient subgraph matching and community detection. This talk will focus on this topic, beginning with a graph neural network-based path dominance embedding method for accurate and error-free exact subgraph matching. Next, it will delve into the dynamic subgraph matching problem, introducing vertex dominance embeddings and degree-aware pruning techniques to efficiently handle subgraph queries on dynamic graphs. Finally, it will present GCLS², a structure-aware graph contrastive learning framework for community detection, which captures structural semantics to achieve accurate and efficient community discovery. The talk will conclude with a brief discussion on future research directions in graph query optimization.

Program Co-Chairs

• Dr. Xiaofei Zhang, University of Memphis

• Dr. Xiang Lian, Kent State University

• Dr. Qiang Guan, Kent State University

Program Committee Members

Instructions for Submission

1. Papers should be formatted to IEEE Computer Society Proceedings Manuscript Formatting Guidelines (https://www.ieee.org/conferences/publishing/templates.html). Papers should be up to 10 pages (references included), in the IEEE 2-column format.
2. You are strongly encouraged to print and double-check your PDF file before its submission, especially if your paper contains Asian/European language symbols (such as Chinese/Korean characters or English letters with European fonts).
3. Please submit through the Cyberchair system, at the following address: https://wi-lab.com/cyberchair/2024/bigdata24/scripts/submit.php?subarea=S20&undisplay_detail=1&wh=/cyberchair/2024/bigdata24/scripts/ws_submit.php
4. The workshop adopts a single-blind review policy.

Contact

For questions or inquiries, please email us at xiaofei.zhang@memphis.edu.