Author name cluster

Wenjing Wang

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

10 papers

1 author row

EAAI Journal 2025 Journal Article

An efficient framework for general long-horizon time series forecasting with Mamba and Diffusion Probabilistic Models

Wenjing Wang
Qilei Li
Ziwu Jiang
Deqian Fu
David Camacho

Time series forecasting plays an essential role in supporting critical decision-making processes in risk management and resource allocation in various fields, including finance, transportation, industrial systems, etc. Conventional models can effectively capture volatility and, are proficient in handling specific patterns, such as the AutoRegressive Integrated Moving Average model (ARIMA) and the Generalized AutoRegressive Conditional Heteroskedasticity model (GARCH). Nonetheless, these models meet many challenges, such as high dimensionality, non-stationarity, and nonlinearity inherent in real-world data. Although deep learning methodologies can provide better performance, they may still suffer from long-term errors and heightened computational expenses. A novel framework named Mamba Diffusion Probabilistic Models (MambaDiffTS) is proposed, which integrates Mamba’s state space model with a frequency-aware diffusion process grounded in Denoising Diffusion Probabilistic Models (DDPM). Mamba’s selective state transitions enable linear-time modeling of long-range dependencies; at the same time, frequency-aware spectral decomposition isolates trends and seasonality through Fourier regularization. Furthermore, the implementation of spectral energy-guided noise scheduling preserves temporal fidelity. Extensive experiments on diverse benchmarks-financial volatility, industrial IoT sensor data, and climate modeling-demonstrate MambaDiffTS’s superiority. Notably, on stock forecasting tasks, MambaDiffTS reduces Mean Squared Error (MSE) by approximately 18. 6% compared to the best-performing baseline, and substantially outperforms diffusion models, all while maintaining linear computational complexity. The proposed MambaDiffTS facilitates scalable forecasting over extended horizons.

Details DOI

AAAI Conference 2024 Conference Paper

Open-Set Facial Expression Recognition

Yuhang Zhang
Yue Yao
Xuannan Liu
Lixiong Qin
Wenjing Wang
Weihong Deng

Facial expression recognition (FER) models are typically trained on datasets with a fixed number of seven basic classes. However, recent research works (Cowen et al. 2021; Bryant et al. 2022; Kollias 2023) point out that there are far more expressions than the basic ones. Thus, when these models are deployed in the real world, they may encounter unknown classes, such as compound expressions that cannot be classified into existing basic classes. To address this issue, we propose the open-set FER task for the first time. Though there are many existing open-set recognition methods, we argue that they do not work well for open-set FER because FER data are all human faces with very small inter-class distances, which makes the open-set samples very similar to close-set samples. In this paper, we are the first to transform the disadvantage of small inter-class distance into an advantage by proposing a new way for open-set FER. Specifically, we find that small inter-class distance allows for sparsely distributed pseudo labels of open-set samples, which can be viewed as symmetric noisy labels. Based on this novel observation, we convert the open-set FER to a noisy label detection problem. We further propose a novel method that incorporates attention map consistency and cycle training to detect the open-set samples. Extensive experiments on various FER datasets demonstrate that our method clearly outperforms state-of-the-art open-set recognition methods by large margins. Code is available at https://github.com/zyh-uaiaaaa.

PDF Details DOI

EAAI Journal 2023 Journal Article

Development of a cross-scale weighted feature fusion network for hot-rolled steel surface defect detection

Yuzhong Zhang
Wenjing Wang
Zhaoming Li
Shuangbao Shu
Xianli Lang
Tengda Zhang
Jingtao Dong

Surface defects of hot-rolled steel would affect the performance and appearance of the final products. In order to detect steel surface defects efficiently, a cross-scale weighted feature fusion network for identifying defect categories and locating defects is proposed in this work. Combined with Laplace sharpening, the backbone in the YOLOv5s model is used to extract multi-scale defect features from input images. And then, an improved weighted bi-directional feature pyramid network embedded with residual modules is proposed to aggregate multi-scale feature maps for enhancing the robustness of multi-size defect representation. Finally, four prediction branches accompanied with prior bounding boxes by a k-means clustering algorithm are responsible for predicting defects with different sizes. The proposed detection network is verified on the NEU-DET dataset, and experimental results show that the proposed network can achieve 86. 8% mAP with the IoU threshold of 0. 5, and can efficiently process images at 51 fps with the RGB image size 640 × 640. The Laplace sharpening module, the k_means clustering module and the improved C3-BiFPN module all contribute to the improvement of performance (mAP) of the proposed network by 1. 8%, 2. 7% and 3. 8%, respectively. Our experimental results demonstrate that the proposed framework can effectively detect the surface defects of hot-rolled steel, and has potential to be used for real-time surface defect detection. Meanwhile, the versatility of the proposed network for other types of defect detection is also evaluated on the MT dataset and the DAGM dataset.

Details DOI

EAAI Journal 2023 Journal Article

Selective Feature Bagging of one-class classifiers for novelty detection in high-dimensional data

Biao Wang
Wenjing Wang
Guanglei Meng
Tiankuo Meng
Bin Song
Yingnan Wang
Yuming Guo
Zhihua Qiao

Novelty detection in high-dimensional data is a challenging task due to the masking effect of irrelevant attributes. A common solution is to discover feature subspace, of which attributes are relevant to novelties. Due to the high uncertainty of novelties in practical applications, ensemble models that combine results from multiple subspaces are proved to be more effective than single models. According to the theory of bias–variance tradeoff, existing ensembles are often developed based on variance reduction. However, it is argued that the combination of poor detectors will deteriorate the performance of ensembles. To this end, this paper proposes an ensemble detector that takes into account variance and bias reduction simultaneously. Our ensemble is referred to as Selective Feature Bagging (SFB) since it is developed on the basis of Feature Bagging (FB). In order to improve the accuracy without deterioration of diversity of base detectors in FB, we resort to the notion of dynamic classifier selection which is proved be effective in classification. During the ensemble generation phase, base detectors are produced and categorized into different groups that are distinguished by the dimensionality of subspace used for training. The purpose of such a design is to maintain the diversity. During the generation phase, the most competent base detector from each of groups is dynamically selected and used to make decision on the test pattern. The purpose of such a design is to enhance the accuracy. We verify the effectiveness of SFB on 15 data sets from KEEL repository. Experimental results have shown that SFB can statistically outperform FB. In addition, several state-of-the-art have also been outperformed by SFB.

Details DOI

EAAI Journal 2022 Journal Article

Boosting the prediction of molten steel temperature in ladle furnace with a dynamic outlier ensemble

Biao Wang
Wenjing Wang
Guanglei Meng
Zhihua Qiao
Yuming Guo
Na Wang
Wei Wang
Zhizhong Mao

Molten steel temperature prediction is a critical step in the development of level-two control systems for ladle furnace. Many machine learning algorithms have been employed to complete such a work. Whereas data-driven predictors often deteriorate due to the presence of outliers in practical applications. This paper proposes to boost the predictive performance via outlier detection. Specifically, a dynamic outlier ensemble is developed inspired by the superiority of dynamic classifier selection in classification. Clustering analysis is used to determine the region of competence, on which base detectors are selected with the dedicated measure. The reason for the usage of clustering analysis lies in its efficiency during online detection. One attribute weighting algorithm is used to enhance the capability of clustering in outlier detection. The information behind regression is used to facilitate the measure of competence, results of which can promote the performance of predictors. Such a strategy can achieve double-win from the perspective of regression and outlier detection. Extensive experiments on real-world data sets show that results of all 4 predictive models with respect to accuracy and hit rate can be improved. Moreover, the detection performance in terms of G-mean and F1 score of our detector has also been confirmed via the comparison with 8 competitors.

Details DOI

EAAI Journal 2022 Journal Article

Dynamic selective Gaussian process regression for forecasting temperature of molten steel in ladle furnace

Biao Wang
Wenjing Wang
Zhihua Qiao
Guanglei Meng
Zhizhong Mao

The requirement for intelligent steelmaking has underlined the significance of data-driven predictions of molten steel temperature in ladle furnace. Recently, predictors based on ensemble learning have shown their superiority over single ones. However, the strong reliability on the ensemble diversity can hardly insure their generalization ability. Moreover, most existing predictors cannot provide statistical meaning to their outputs. This has degraded their engineering value. In this paper, we aim to address these two problems in one scheme, where a dynamic regression ensemble of Gaussian process models is built. Our dynamic ensemble will select the most competent individual for each test pattern according to the competence estimated by informative neighbors. To this end, a distance measure based on RReliefF is constructed to search for these neighbors, rather than traditional K-nearest neighbor. Several evaluation indexes are combined by a meta regressor so that more robust estimation of competence can be achieved. A Bayesian nonparametric model is used for ensemble generation in order to obtain statistical predictions. A data set from real-world ladle furnace is used to verify the effectiveness of the proposed predictor. According to the comparative results, we have found the superiority of our dynamic ensemble over static ensembles and single predictors. Furthermore, the improvement over existing dynamic ensembles has also been confirmed.

Details DOI

IJCAI Conference 2021 Conference Paper

Instance-Aware Coherent Video Style Transfer for Chinese Ink Wash Painting

Hao Liang
Shuai Yang
Wenjing Wang
Jiaying Liu

Recent researches have made remarkable achievements in fast video style transfer based on western paintings. However, due to the inherent different drawing techniques and aesthetic expressions of Chinese ink wash painting, existing methods either achieve poor temporal consistency or fail to transfer the key freehand brushstroke characteristics of Chinese ink wash painting. In this paper, we present a novel video style transfer framework for Chinese ink wash paintings. The two key ideas are a multi-frame fusion for temporal coherence and an instance-aware style transfer. The frame reordering and stylization based on reference frame fusion are proposed to improve temporal consistency. Meanwhile, the proposed method is able to adaptively leave the white spaces in the background and to select proper scales to extract features and depict the foreground subject by leveraging instance segmentation. Experimental results demonstrate the superiority of the proposed method over state-of-the-art style transfer methods in terms of both temporal coherence and visual quality. Our project website is available at https: //oblivioussy. github. io/InkVideo/.

PDF Details DOI

IJCAI Conference 2021 Conference Paper

Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification

Yunsheng Shi
Zhengjie Huang
Shikun Feng
Hui Zhong
Wenjing Wang
Yu Sun

Graph neural network (GNN) and label propagation algorithm (LPA) are both message passing algorithms, which have achieved superior performance in semi-supervised classification. GNN performs feature propagation by a neural network to make predictions, while LPA uses label propagation across graph adjacency matrix to get results. However, there is still no effective way to directly combine these two kinds of algorithms. To address this issue, we propose a novel Unified Message Passaging Model (UniMP) that can incorporate feature and label propagation at both training and inference time. First, UniMP adopts a Graph Transformer network, taking feature embedding and label embedding as input information for propagation. Second, to train the network without overfitting in self-loop input label information, UniMP introduces a masked label prediction strategy, in which some percentage of input label information are masked at random, and then predicted. UniMP conceptually unifies feature propagation and label propagation and is empirically powerful. It obtains new state-of-the-art semi-supervised classification results in Open Graph Benchmark (OGB).

PDF Details DOI

AAAI Conference 2020 Conference Paper

Consistent Video Style Transfer via Compound Regularization

Wenjing Wang
Jizheng Xu
Li Zhang
Yue Wang
Jiaying Liu

Recently, neural style transfer has drawn many attentions and signiﬁcant progresses have been made, especially for image style transfer. However, ﬂexible and consistent style transfer for videos remains a challenging problem. Existing training strategies, either using a signiﬁcant amount of video data with optical ﬂows or introducing single-frame regularizers, have limited performance on real videos. In this paper, we propose a novel interpretation of temporal consistency, based on which we analyze the drawbacks of existing training strategies; and then derive a new compound regularization. Experimental results show that the proposed regularization can better balance the spatial and temporal performance, which supports our modeling. Combining with the new cost formula, we design a zero-shot video style transfer framework. Moreover, for better feature migration, we introduce a new module to dynamically adjust inter-channel distributions. Quantitative and qualitative results demonstrate the superiority of our method over other state-of-the-art style transfer methods. Our project is publicly available at: https: //daooshee. github. io/CompoundVST/.

PDF Details

AAAI Conference 2019 Conference Paper

TET-GAN: Text Effects Transfer via Stylization and Destylization

Shuai Yang
Jiaying Liu
Wenjing Wang
Zongming Guo

Text effects transfer technology automatically makes the text dramatically more impressive. However, previous style transfer methods either study the model for general style, which cannot handle the highly-structured text effects along the glyph, or require manual design of subtle matching criteria for text effects. In this paper, we focus on the use of the powerful representation abilities of deep neural features for text effects transfer. For this purpose, we propose a novel Texture Effects Transfer GAN (TET-GAN), which consists of a stylization subnetwork and a destylization subnetwork. The key idea is to train our network to accomplish both the objective of style transfer and style removal, so that it can learn to disentangle and recombine the content and style features of text effects images. To support the training of our network, we propose a new text effects dataset with as much as 64 professionally designed styles on 837 characters. We show that the disentangled feature representations enable us to transfer or remove all these styles on arbitrary glyphs using one network. Furthermore, the flexible network design empowers TET-GAN to efficiently extend to a new text style via oneshot learning where only one example is required. We demonstrate the superiority of the proposed method in generating high-quality stylized text over the state-of-the-art methods.

PDF Details