Dachun Sun
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Dachun Sun has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Tarek Abdelzaher
University of Illinois at Urbana-Champaign, Champaign, USA
9 shared publications - Ruijie Wang
School of Computer Science and Engineering, Beihang University, Beijing, China
3 shared publications - Xinyi Liu
University of Illinois Urbana-Champaign, Champaign, IL, USA
3 shared publications - Boleslaw Karol Szymanski
Rensselaer Polytechnic Institute
2 shared publications - Dilek Hakkani-Tür
University of Illinois Urbana-Champaign, Urbana, IL, USA
2 shared publications - Huajie Shao
Department of Computer Science, William & Mary, Williamsburg, Virginia, USA
2 shared publications
Latest Publications
- 3 February 2026
- Published by Springer Nature
- p. 230-247
GPSocio: A Transformer-Based General-Purpose Social Network Representation System
- 3 February 2026
- Published by Springer Nature
- p. 230-247
GPSocio: A Transformer-Based General-Purpose Social Network Representation System
- 26 January 2026
- Published by Springer Nature
- p. 299-314
Beliefs in Motion: Simulating Opinion Dynamics via LLM-Powered Community Reactions
- 26 January 2026
- Published by Springer Nature
- p. 299-314
Beliefs in Motion: Simulating Opinion Dynamics via LLM-Powered Community Reactions
- Open Access
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 1388-1399
Learning to Slice: Self-Supervised Interpretable Hierarchical Representation Learning with Graph Auto-Encoder Tree
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 1388-1399
Open AccessLearning to Slice: Self-Supervised Interpretable Hierarchical Representation Learning with Graph Auto-Encoder Tree
- Open Access
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2736-2743
Uncovering Cross-Domain Recommendation Ability of Large Language Models
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2736-2743
Open AccessUncovering Cross-Domain Recommendation Ability of Large Language Models
- Open Access
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 5530-5533
Towards Efficient Temporal Graph Learning: Algorithms, Frameworks, and Tools
Temporal graphs capture dynamic node relations via temporal edges, finding extensive utility in wide domains where time-varying patterns are crucial. Temporal Graph Neural Networks (TGNNs) have gained significant attention for their effectiveness in representing temporal graphs. However, TGNNs still face significant efficiency challenges in real-world low-resource settings. First, from a data-efficiency standpoint, training TGNNs requires sufficient temporal edges and data labels, which is problematic in practical scenarios with limited data collection and annotation. Second, from a resource-efficiency perspective, TGNN training and inference are computationally demanding due to complex encoding operations, especially on large-scale temporal graphs. Minimizing resource consumption while preserving effectiveness is essential. Inspired by these efficiency challenges, this tutorial systematically introduces state-of-the-art data-efficient and resource-efficient TGNNs, focusing on algorithms, frameworks, and tools, and discusses promising yet under-explored research directions in efficient temporal graph learning. This tutorial aims to benefit researchers and practitioners in data mining, machine learning, and artificial intelligence.
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 5530-5533
Open AccessTowards Efficient Temporal Graph Learning: Algorithms, Frameworks, and Tools
Temporal graphs capture dynamic node relations via temporal edges, finding extensive utility in wide domains where time-varying patterns are crucial. Temporal Graph Neural Networks (TGNNs) have gained significant attention for their effectiveness in representing temporal graphs. However, TGNNs still face significant efficiency challenges in real-world low-resource settings. First, from a data-efficiency standpoint, training TGNNs requires sufficient temporal edges and data labels, which is problematic in practical scenarios with limited data collection and annotation. Second, from a resource-efficiency perspective, TGNN training and inference are computationally demanding due to complex encoding operations, especially on large-scale temporal graphs. Minimizing resource consumption while preserving effectiveness is essential. Inspired by these efficiency challenges, this tutorial systematically introduces state-of-the-art data-efficient and resource-efficient TGNNs, focusing on algorithms, frameworks, and tools, and discusses promising yet under-explored research directions in efficient temporal graph learning. This tutorial aims to benefit researchers and practitioners in data mining, machine learning, and artificial intelligence.
- Open Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 59-69
MetaHKG: Meta Hyperbolic Learning for Few-shot Temporal Reasoning
This paper investigates the few-shot temporal reasoning capability within the hyperbolic space. The goal is to forecast future events for newly emerging entities within temporal knowledge graphs (TKGs), leveraging only a limited set of initial observations. Hyperbolic space is advantageous for modeling emerging graph entities for two reasons: First, its geometric property of exponential expansion aligns with the rapid growth of new entities in real-world graphs; Second, it excels in capturing power-law patterns and hierarchical structures, well-suitable for new entities distributed at the peripheries of graph hierarchies and loosely connected with others through few links. We therefore propose a meta-learning framework, MetaHKG, to enable few-shot temporal reasoning within a hyperbolic space. Unlike prior hyperbolic learning works, MetaHKG addresses the challenges of effectively representing new entities in TKGs and adapting model parameters by incorporating novel hyperbolic time encodings and temporal attention networks that achieve translational invariance. We also introduce a meta hyperbolic optimization algorithm to enhance model adaptation by learning both global and entity-specific parameters through bi-level optimization. Comprehensive experiments conducted on three real-world temporal knowledge graphs demonstrate the superiority of MetaHKG over a diverse range of baselines, which achieves average 5.2% relative improvements. Compared to its Euclidean counterpart, MetaHKG operates in a lower-dimensional space but yields a more stable and efficient adaptability towards new entities.
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 59-69
Open AccessMetaHKG: Meta Hyperbolic Learning for Few-shot Temporal Reasoning
This paper investigates the few-shot temporal reasoning capability within the hyperbolic space. The goal is to forecast future events for newly emerging entities within temporal knowledge graphs (TKGs), leveraging only a limited set of initial observations. Hyperbolic space is advantageous for modeling emerging graph entities for two reasons: First, its geometric property of exponential expansion aligns with the rapid growth of new entities in real-world graphs; Second, it excels in capturing power-law patterns and hierarchical structures, well-suitable for new entities distributed at the peripheries of graph hierarchies and loosely connected with others through few links. We therefore propose a meta-learning framework, MetaHKG, to enable few-shot temporal reasoning within a hyperbolic space. Unlike prior hyperbolic learning works, MetaHKG addresses the challenges of effectively representing new entities in TKGs and adapting model parameters by incorporating novel hyperbolic time encodings and temporal attention networks that achieve translational invariance. We also introduce a meta hyperbolic optimization algorithm to enhance model adaptation by learning both global and entity-specific parameters through bi-level optimization. Comprehensive experiments conducted on three real-world temporal knowledge graphs demonstrate the superiority of MetaHKG over a diverse range of baselines, which achieves average 5.2% relative improvements. Compared to its Euclidean counterpart, MetaHKG operates in a lower-dimensional space but yields a more stable and efficient adaptability towards new entities.
- 6 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Control of Network Systems
- Vol. 11 (2) , 964-976
The Paradox of Information Access: On Modeling Polarization in the Age of Information
The paper derives a new nonlinear stochastic model of evolution of human beliefs that demonstrates how an increase in democratized information production and sharing, combined with consumers' confirmation bias and natural bias for outlying content, result in increased polarization. The model shows that the evolution of human beliefs can be approximated by a nonlinear diffusion-drift equation in which systematic psychological biases contribute to drift , whereas other random influences contribute to diffusion . The nonlinear formulation predicts a growth in polarization that is attributable to increasing information production and sharing . While the core contribution is analytical, an anecdotal model parameter fitting to empirical data is also presented. Specifically, we show that our model closely predicts the changing and increasingly polarized distribution of ideology of members of the US Congress over the last quarter-century (taken as an approximate proxy for shifts in the US population ideology), when we take the mobile phone penetration curve as a proxy for democratization of information access. The model suggests that escaping the polarizing forces in the age of information access may be an uphill battle.
- 6 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Control of Network Systems
- Vol. 11 (2) , 964-976
The Paradox of Information Access: On Modeling Polarization in the Age of Information
The paper derives a new nonlinear stochastic model of evolution of human beliefs that demonstrates how an increase in democratized information production and sharing, combined with consumers' confirmation bias and natural bias for outlying content, result in increased polarization. The model shows that the evolution of human beliefs can be approximated by a nonlinear diffusion-drift equation in which systematic psychological biases contribute to drift , whereas other random influences contribute to diffusion . The nonlinear formulation predicts a growth in polarization that is attributable to increasing information production and sharing . While the core contribution is analytical, an anecdotal model parameter fitting to empirical data is also presented. Specifically, we show that our model closely predicts the changing and increasingly polarized distribution of ideology of members of the US Congress over the last quarter-century (taken as an approximate proxy for shifts in the US population ideology), when we take the mobile phone penetration curve as a proxy for democratization of information access. The model suggests that escaping the polarizing forces in the age of information access may be an uphill battle.
- Open Access
- 22 December 2021
- Published by Frontiers Media SA in Frontiers in Big Data
- Vol. 4, 729881
Computational Modeling of Hierarchically Polarized Groups by Structured Matrix Factorization
The paper extends earlier work on modeling hierarchically polarized groups on social media. An algorithm is described that 1) detects points ofagreementanddisagreementbetween groups, and 2) divides themhierarchicallyto represent nested patterns of agreement and disagreement given a structural guide. For example, two opposing parties might disagree on core issues. Moreover, within a party, despite agreement on fundamentals, disagreement might occur on further details. We call such scenarioshierarchically polarized groups. An (enhanced) unsupervised Non-negative Matrix Factorization (NMF) algorithm is described for computational modeling of hierarchically polarized groups. It is enhanced with a language model, and with a proof of orthogonality of factorized components. We evaluate it on both synthetic and real-world datasets, demonstrating ability to hierarchically decompose overlapping beliefs. In the case where polarization is flat, we compare it to prior art and show that it outperforms state of the art approaches for polarization detection and stance separation. An ablation study further illustrates the value of individual components, including new enhancements.
- 22 December 2021
- Published by Frontiers Media SA in Frontiers in Big Data
- Vol. 4, 729881
Open AccessComputational Modeling of Hierarchically Polarized Groups by Structured Matrix Factorization
The paper extends earlier work on modeling hierarchically polarized groups on social media. An algorithm is described that 1) detects points ofagreementanddisagreementbetween groups, and 2) divides themhierarchicallyto represent nested patterns of agreement and disagreement given a structural guide. For example, two opposing parties might disagree on core issues. Moreover, within a party, despite agreement on fundamentals, disagreement might occur on further details. We call such scenarioshierarchically polarized groups. An (enhanced) unsupervised Non-negative Matrix Factorization (NMF) algorithm is described for computational modeling of hierarchically polarized groups. It is enhanced with a language model, and with a proof of orthogonality of factorized components. We evaluate it on both synthetic and real-world datasets, demonstrating ability to hierarchically decompose overlapping beliefs. In the case where polarization is flat, we compare it to prior art and show that it outperforms state of the art approaches for polarization detection and stance separation. An ablation study further illustrates the value of individual components, including new enhancements.
- 17 November 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (12) , 9285-9297
ControlVAE: Tuning, Analytical Properties, and Performance Analysis
This paper reviews the novel concept of a controllable variational autoencoder (ControlVAE), discusses its parameter tuning to meet application needs, derives its key analytic properties, and offers useful extensions and applications. ControlVAE is a new variational autoencoder (VAE) framework that combines automatic control theory with the basic VAE to stabilize the KL-divergence of VAE models to a specified value. It leverages a non-linear PI controller, a variant of the proportional-integral-derivative (PID) controller, to dynamically tune the weight of the KL-divergence term in the evidence lower bound (ELBO) using the output KL-divergence as feedback. This allows us to precisely control the KL-divergence to a desired value (set point) that is effective in avoiding posterior collapse and learning disentangled representations. While prior work developed alternative techniques for controlling the KL divergence, we show that our PI controller has better stability properties and thus better convergence, thereby producing better disentangled representations from finite training data. In order to improve the ELBO of ControlVAE over that of the regular VAE, we provide a simplified theoretical analysis to inform the choice of set point for the KL-divergence of ControlVAE. We evaluate the proposed method on three tasks: image generation, language modeling, and disentangled representation learning. The results show that ControlVAE can achieve much better reconstruction quality than the other methods for comparable disentanglement. On the language modeling task, our method can avoid posterior collapse (KL vanishing) and improve the diversity of generated text. Moreover, it can change the optimization trajectory, improving the ELBO and the reconstruction quality for image generation.
- 17 November 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (12) , 9285-9297
ControlVAE: Tuning, Analytical Properties, and Performance Analysis
This paper reviews the novel concept of a controllable variational autoencoder (ControlVAE), discusses its parameter tuning to meet application needs, derives its key analytic properties, and offers useful extensions and applications. ControlVAE is a new variational autoencoder (VAE) framework that combines automatic control theory with the basic VAE to stabilize the KL-divergence of VAE models to a specified value. It leverages a non-linear PI controller, a variant of the proportional-integral-derivative (PID) controller, to dynamically tune the weight of the KL-divergence term in the evidence lower bound (ELBO) using the output KL-divergence as feedback. This allows us to precisely control the KL-divergence to a desired value (set point) that is effective in avoiding posterior collapse and learning disentangled representations. While prior work developed alternative techniques for controlling the KL divergence, we show that our PI controller has better stability properties and thus better convergence, thereby producing better disentangled representations from finite training data. In order to improve the ELBO of ControlVAE over that of the regular VAE, we provide a simplified theoretical analysis to inform the choice of set point for the KL-divergence of ControlVAE. We evaluate the proposed method on three tasks: image generation, language modeling, and disentangled representation learning. The results show that ControlVAE can achieve much better reconstruction quality than the other methods for comparable disentanglement. On the language modeling task, our method can avoid posterior collapse (KL vanishing) and improve the diversity of generated text. Moreover, it can change the optimization trajectory, improving the ELBO and the reconstruction quality for image generation.
- 10 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 70 (9) , 1325-1337
Truth Discovery With Multi-Modal Data in Social Sensing
This paper proposes unsupervised truth-finding algorithms that combine consideration of multi-modal content features with analysis of propagation patterns to evaluate the veracity of observations in social sensing applications. A key challenge of social sensing is to develop effective algorithms for estimating both the reliability of sources and the veracity of their observations without prior knowledge. In contrast to prior solutions that use labeled examples to learn content features, our approach is entirely unsupervised. Hence, given no prior training data, we jointly learn the importance of different content features together with the veracity of observations using propagation patterns as an indicator of perceived content reliability. Consider the correlation between different features, a novel penalized expectation maximization (PEM) algorithm is proposed to improve the quality of estimation results for observations bolstered by multiple features. In addition, we further develop a constrained EM with multiple features (CEM-MultiF) that incorporates multiple corroborating features as a constraint to boost the estimation accuracy. Finally, we evaluate the performance of the proposed algorithms on real-world data sets collected from Twitter. The evaluation results demonstrate that the proposed algorithms outperform the existing fact-finding approaches, and offer tunable knobs for controlling robustness/performance trade-offs in the presence of malicious sources.
- 10 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 70 (9) , 1325-1337
Truth Discovery With Multi-Modal Data in Social Sensing
This paper proposes unsupervised truth-finding algorithms that combine consideration of multi-modal content features with analysis of propagation patterns to evaluate the veracity of observations in social sensing applications. A key challenge of social sensing is to develop effective algorithms for estimating both the reliability of sources and the veracity of their observations without prior knowledge. In contrast to prior solutions that use labeled examples to learn content features, our approach is entirely unsupervised. Hence, given no prior training data, we jointly learn the importance of different content features together with the veracity of observations using propagation patterns as an indicator of perceived content reliability. Consider the correlation between different features, a novel penalized expectation maximization (PEM) algorithm is proposed to improve the quality of estimation results for observations bolstered by multiple features. In addition, we further develop a constrained EM with multiple features (CEM-MultiF) that incorporates multiple corroborating features as a constraint to boost the estimation accuracy. Finally, we evaluate the performance of the proposed algorithms on real-world data sets collected from Twitter. The evaluation results demonstrate that the proposed algorithms outperform the existing fact-finding approaches, and offer tunable knobs for controlling robustness/performance trade-offs in the presence of malicious sources.