Songgaojun Deng

Songgaojun Deng

Auto-generated profile, based on publicly available information

Songgaojun Deng

AIRLab, University of Amsterdam, Amsterdam, The Netherlands

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 November 2023 in Knowledge and Information Systems
Verified
Journal Article
Closed access
PDF file
CLCheng Liu
Cheng Liu
CYChenhuan Yu
Chenhuan Yu
NGNing Gui
Ning Gui
ZYZhiwu Yu
Zhiwu Yu
SDSonggaojun Deng
Songgaojun Deng

27 February 2024 in ACM Transactions on Knowledge Discovery From Data
Verified
Journal Article
Open access
SDSonggaojun Deng
Songgaojun Deng
OSOlivier Sprangers
Olivier Sprangers
MLMing Li
Ming Li
SSSebastian Schelter
Sebastian Schelter
MRMaarten de Rijke
Maarten de Rijke
Domain generalization aims to design models that can effectively generalize to unseen target domains by learning from observed source domains. Domain generalization poses a significant challenge for time series data, due to varying data distributions and temporal dependencies. Existing approaches to domain generalization are not designed for time series data, which often results in suboptimal or unstable performance when confronted with diverse temporal patterns and complex data characteristics. We propose a novel approach to tackle the problem of domain generalization in time series forecasting. We focus on a scenario where time series domains share certain common attributes and exhibit no abrupt distribution shifts. Our method revolves around the incorporation of a key regularization term into an existing time series forecasting model:domain discrepancy regularization. In this way, we aim to enforce consistent performance across different domains that exhibit distinct patterns. We calibrate the regularization term by investigating the performance within individual domains and propose thedomain discrepancy regularization with domain difficulty awareness. We demonstrate the effectiveness of our method on multiple datasets, including synthetic and real-world time series datasets from diverse domains such as retail, transportation, and finance. Our method is compared against traditional methods, deep learning models, and domain generalization approaches to provide comprehensive insights into its performance. In these experiments, our method showcases superior performance, surpassing both the base model and competing domain generalization models across all datasets. Furthermore, our method is highly general and can be applied to various time series models.

24 August 2024
Verified
Conference Paper
Open access
SDSonggaojun Deng
Songgaojun Deng
MRMaarten de Rijke
Maarten de Rijke
YNYue Ning
Yue Ning
Human events such as hospital visits, protests, and epidemic outbreaks directly affect individuals, communities, and societies. These events are often influenced by factors such as economics, politics, and public policies of our society. The abundance of online data sources such as social networks, official news articles, and personal blogs chronicle societal events, facilitating the development of AI models for social science, public health care, and decision making. Human event modeling generally comprises both the forecasting stage, which estimates future events based on historical data, and interpretation, which seeks to identify influential factors of such events to understand their causative attributes. Recent achievements, fueled by deep learning and the availability of public data, have significantly advanced the field of human event modeling. This survey offers a systematic overview of deep learning technologies for forecasting and interpreting human events, with a primary focus on political events. We first introduce the existing challenges and background in this domain. We then present the problem formulation of event forecasting and interpretation. We investigate recent achievements in graph neural networks, owing to the prevalence of relational data and the efficacy of graph learning models. We also discuss the latest studies that utilize large language models for event reasoning. Lastly, we provide summaries of data resources, open challenges, and future research directions in the study of human event modeling.

13 December 2018 in Genes
Verified
Journal Article
Open access
PDF file
XYXu Yang
Xu Yang
SDSonggaojun Deng
Songgaojun Deng
MJMengyao Ji
Mengyao Ji
JZJinfeng Zhao
Jinfeng Zhao
WZWenhao Zheng
Wenhao Zheng
Artificial intelligence research received more and more attention nowadays. Neural Evolution (NE) is one very important branch of AI, which waves the power of evolutionary algorithms to generate Artificial Neural Networks (ANNs). How to use the evolutionary advantages of network topology and weights to solve the application of Artificial Neural Networks is the main problem in the field of NE. In this paper, a novel DNA encoding method based on the triple codon is proposed. Additionally, a NE algorithm Triplet Codon Encoding Neural Network Evolving Algorithm (TCENNE) based on this encoding method is presented to verify the rationality and validity of the coding design. The results show that TCENNE is very effective and more robust than NE algorithms, due to the coding design. Also, it is shown that it can realize the co-evolution of network topology and weights and outperform other neural evolution systems in challenging reinforcement learning tasks.

30 September 2024 in arXiv
Verified
Preprint
Open access
PDF file
WYWeiwei Ye
Weiwei Ye
SDSonggaojun Deng
Songgaojun Deng
QZQiaosha Zou
Qiaosha Zou
NGNing Gui
Ning Gui
Time series forecasting typically needs to address non-stationary data with evolving trend and seasonal patterns. To address the non-stationarity, reversible instance normalization has been recently proposed to alleviate impacts from the trend with certain statistical measures, e.g., mean and variance. Although they demonstrate improved predictive accuracy, they are limited to expressing basic trends and are incapable of handling seasonal patterns. To address this limitation, this paper proposes a new instance normalization solution, called frequency adaptive normalization (FAN), which extends instance normalization in handling both dynamic trend and seasonal patterns. Specifically, we employ the Fourier transform to identify instance-wise predominant frequent components that cover most non-stationary factors. Furthermore, the discrepancy of those frequency components between inputs and outputs is explicitly modeled as a prediction task with a simple MLP model. FAN is a model-agnostic method that can be applied to arbitrary predictive backbones. We instantiate FAN on four widely used forecasting models as the backbone and evaluate their prediction performance improvements on eight benchmark datasets. FAN demonstrates significant performance advancement, achieving 7.76% ~ 37.90% average improvements in MSE.

18 June 2026 in ACM Transactions on Knowledge Discovery From Data
Verified
Journal Article
Open access
SDSonggaojun Deng
Songgaojun Deng
ZXZehao Xiao
Zehao Xiao
MRMaarten de Rijke
Maarten de Rijke
Time series forecasting is essential in many real-world applications, yet developing models that generalize well to unseen and related domains – such as forecasting web traffic on new websites/platforms or predicting e-commerce demand in new regions – remains a big challenge. Prior work addressing this problem, known as domain generalization, focuses on identifying common patterns but often overlooks the complex characteristics of time series data and fails to use the available information from test samples of unseen domains. We propose a novel approach, adaptive latent decomposition (ALD) for domain generalization in time series forecasting, which consists of a decomposed variational autoencoder (VAE) and an adaptive inference mechanism to improve predictive performance in unseen domains. The decomposed VAE involves a learnable kernel-selection mechanism that learns latent variables of decomposed components of time series, i.e., trend-cyclical and seasonal components. These latent variables capture hidden temporal dependencies in time series data, allowing forecasting models to learn general patterns from various seen domains. The adaptive inference mechanism bridges the gap between seen and unseen domains with a sample-wise optimization strategy specifically designed for time series forecasting. ALD builds a latent variables-aware pre-trained model and tailors it for each test sample, improving the generalization on unseen test domains. We validate ALD across six real-world datasets, from online behavior to complex temporal systems, demonstrating its superior generalization performance compared to state-of-the-art methods.