Author name cluster

Tianyang 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.

9 papers

1 author row

EAAI Journal 2026 Journal Article

A novel interpretable dynamic weighted domain adaptation network for cross-domain fault diagnosis of bearings under time-varying speeds

Xueyi Li
Sixin Li
Guangyao Zhang
Yining Xie
Tianyang Wang
Fulei Chu

Significant progress has been made in rolling bearing fault diagnosis based on unsupervised domain adaptation (UDA). However, in complex operating conditions, especially under time-varying speed conditions, these methods still face critical challenges, including severe distribution discrepancies between source and target domains and limited interpretability of the diagnostic process. To address these issues, this paper proposes a novel interpretable dynamic weighted domain adaptation network (DWDAN), which combines a discrete wavelet-guided attention (DW-GA) layer with dynamic weighted joint domain adaptation (DWJDA) to achieve subdomain-level alignment while enhancing diagnostic interpretability. Specifically, the DW-GA layer incorporates physical prior knowledge to guide the model in extracting fault-related features in the wavelet domain, thereby improving the efficiency and effectiveness of feature learning under time-varying speed conditions. The DWJDA further performs a refined dynamic adjustment of marginal and conditional distribution alignment by jointly considering intra-class sample weighting and the relative importance of different distribution discrepancies, leading to enhanced domain adaptation capability. The comparative experiments are conducted on the datasets of Huazhong University of Science and Technology (HUST) and Northeast Forestry University (NEFU). The experimental results show that, compared with other mainstream methods, DWDAN exhibits significant advantages in cross-domain fault diagnosis tasks under time-varying speeds, achieving a maximum average diagnostic accuracy of 99. 08%. Further ablation experiments further verify the effectiveness of each key module in improving both diagnostic performance and interpretability, indicating that DWDAN has strong application potential for bearing fault diagnosis under time-varying speed.

Details DOI

AAAI Conference 2026 Conference Paper

CAD-VAE: Leveraging Correlation-Aware Latents for Comprehensive Fair Disentanglement

Chenrui Ma
Xi Xiao
Tianyang Wang
Xiao Wang
Yanning Shen

While deep generative models have significantly advanced representation learning, they may inherit or amplify biases and fairness issues by encoding sensitive attributes alongside predictive features. Enforcing strict independence in disentanglement is often unrealistic when target and sensitive factors are naturally correlated. To address this challenge, we propose CAD-VAE(Correlation-Aware Disentangled VAE), which introduces a correlated latent code to capture the information shared between the target and sensitive attributes. Given this correlated latent, our method effectively separates overlapping factors without extra domain knowledge by directly minimizing the conditional mutual information between target and sensitive codes. A relevance-driven optimization strategy refines the correlated code by efficiently capturing essential correlated features and eliminating redundancy. Extensive experiments on benchmark datasets demonstrate that CAD-VAE produces fairer representations, realistic counterfactuals, and improved fairness-aware image editing.

PDF Details DOI

TMLR Journal 2026 Journal Article

Prompt-based Adaptation in Large-scale Vision Models: A Survey

Xi Xiao
Yunbei Zhang
Lin Zhao
Yiyang Liu
Xiaoying Liao
Zheda Mai
Xingjian Li
Xiao Wang

In computer vision, Visual Prompting (VP) and Visual Prompt Tuning (VPT) have recently emerged as lightweight and effective alternatives to full fine-tuning for adapting large-scale vision models within the ``pretrain-then-finetune'' paradigm. However, despite rapid progress, their conceptual boundaries remain blurred, as VP and VPT are frequently used interchangeably in current research, reflecting a lack of systematic distinction between these techniques and their respective applications. In this survey, we revisit the designs of VP and VPT from first principles, and conceptualize them within a unified framework termed Prompt-based Adaptation (PA). Within this framework, we distinguish methods based on their injection granularity: VP operates at the pixel level, while VPT injects prompts at the token level. We further categorize these methods by their generation mechanism into fixed, learnable, and generated prompts. Beyond the core methodologies, we examine PA’s integrations across diverse domains, including medical imaging, 3D point clouds, and vision-language tasks, as well as its role in test-time adaptation and trustworthy AI. We also summarize current benchmarks and identify key challenges and future directions. To the best of our knowledge, we are the first comprehensive survey dedicated to PA's methodologies and applications in light of their distinct characteristics. Our survey aims to provide a clear roadmap for researchers and practitioners in all area to understand and explore the evolving landscape of PA-related research.

PDF Details

IJCAI Conference 2025 Conference Paper

Faster Annotation for Elevation-Guided Flood Extent Mapping by Consistency-Enhanced Active Learning

Saugat Adhikari
Da Yan
Tianyang Wang
Landon Dyken
Sidharth Kumar
Lyuheng Yuan
Akhlaque Ahmad
Jiao Han

Flood extent mapping is crucial for disaster response and damage assessment. While Earth imagery and terrain data (in the form of DEM) are now readily available, there are few flood annotation data for training machine learning models, which hinders the automated mapping of flooded areas. We propose ALFA, an interactive active-learning-based approach to minimize the annotators' efforts when preparing the ground-truth flood map in a satellite image. ALFA calibrates the prediction consistency of a segmentation model (1) across training cycles and (2) for various data augmentations. The two consistencies are integrated into the design of both the acquisition function and the loss function to enhance the robustness of active learning with limited annotation inputs. ALFA recommends those superpixels that the underlying model is most uncertain about, and users can annotate their pixels with minimal clicks with the help of elevation guidance. Extensive experiments on various regions hit by flooding show that we can improve the annotation time from hours to around 20 minutes. ALFA is open sourced at https: //github. com/saugatadhikari/alfa.

PDF Details DOI

EAAI Journal 2025 Journal Article

Fault diagnosis method for imbalanced data based on adaptive diffusion models and generative adversarial networks

Xueyi Li
Xudong Wu
Tianyang Wang
Yining Xie
Fulei Chu

In engineering practice, the challenge of collecting fault samples often leads to data imbalance, significantly affecting the performance of fault diagnosis. Data augmentation methods offer an effective solution to address this issue by supplementing fault data. However, these methods must meet the requirements of generating high-quality samples, achieving comprehensive mode coverage, and ensuring fast sample generation. Traditional Generative Adversarial Networks are prone to mode collapse, while diffusion models suffer from slow generation speeds, making it difficult for either to fully satisfy these demands. To overcome these challenges, this paper proposes a data augmentation method for fault diagnosis based on an adaptive diffusion model integrated with Generative Adversarial Networks. By introducing a length-adaptive forward diffusion chain to generate Gaussian mixture noise, this method not only ensures smoother and more stable gradients but also avoids computational redundancy and gradient vanishing problems. At each diffusion timestep, the discriminator learns to distinguish real and generated data across different noise ratios and timesteps, enhancing the diversity of fault samples and effectively mitigating mode collapse. Experimental results on two datasets demonstrate that the proposed method outperforms other data augmentation techniques in terms of generation efficiency and stability, effectively addressing the data imbalance problem and significantly improving fault diagnosis performance.

Details DOI

NeurIPS Conference 2025 Conference Paper

MoRE-Brain: Routed Mixture of Experts for Interpretable and Generalizable Cross-Subject fMRI Visual Decoding

Yuxiang Wei
Yanteng Zhang
Xi Xiao
Tianyang Wang
Xiao Wang
Vince D. Calhoun

Decoding visual experiences from fMRI offers a powerful avenue to understand human perception and develop advanced brain-computer interfaces. However, current progress often prioritizes maximizing reconstruction fidelity while overlooking interpretability, an essential aspect for deriving neuroscientific insight. To address this gap, we propose MoRE-Brain, a neuro-inspired framework designed for high-fidelity, adaptable, and interpretable visual reconstruction. MoRE-Brain uniquely employs a hierarchical Mixture-of-Experts architecture where distinct experts process fMRI signals from functionally related voxel groups, mimicking specialized brain networks. The experts are first trained to encode fMRI into the frozen CLIP space. A finetuned diffusion model then synthesizes images, guided by expert outputs through a novel dual-stage routing mechanism that dynamically weighs expert contributions across the diffusion process. MoRE-Brain offers three main advancements: First, it introduces a novel Mixture-of-Experts architecture grounded in brain network principles for neuro-decoding. Second, it achieves efficient cross-subject generalization by sharing core expert networks while adapting only subject-specific routers. Third, it provides enhanced mechanistic insight, as the explicit routing reveals precisely how different modeled brain regions shape the semantic and spatial attributes of the reconstructed image. Extensive experiments validate MoRE-Brain’s high reconstruction fidelity, with bottleneck analyses further demonstrating its effective utilization of fMRI signals, distinguishing genuine neural decoding from over-reliance on generative priors. Consequently, MoRE-Brain marks a substantial advance towards more generalizable and interpretable fMRI-based visual decoding.

PDF Details

EAAI Journal 2025 Journal Article

Multimodal data imputation and fusion for trustworthy fault diagnosis of mechanical systems

Jie Zhang
Yun Kong
Qinkai Han
Tianyang Wang
Mingming Dong
Hui Liu
Fulei Chu

The presence of missing values in the collected data due to sensor failure, communication interruption, or environmental interference can greatly diminishes the trustworthiness of fault diagnosis for mechanical systems. Therefore, this study proposes and evaluates a novel multimodal data imputation and fusion method to perform the trustworthy fault diagnosis of mechanical systems. First, a generative adversarial imputation network, termed as the L2 regularization temporal–spatial generative adversarial imputation network (L2-TSGAIN), is developed. This L2-TSGAIN network, based on a temporal–spatial feature extraction module and L2 regularization loss function, can comprehensively extract data features from both temporal and spatial perspectives, thus achieving high-quality imputation of anomalous sensor data. Subsequently, a multi-input single-output autoencoder (MISO-AE) is designed to extract a universal representation of the imputed data from different modalities and recover features in the fusion data. Finally, the fusion data from different health states of mechanical systems are input into a convolutional neural network classifier to perform fault diagnosis. Experiment validations, considering the presence of missing values in sensor data, have been carried out on the planetary transmission system and gearbox test bench. Compared with several mainstream data imputation methods for fault diagnosis, the optimal diagnostic accuracy of 99. 68 % and 100 % on these two datasets can be obtained using the proposed method, respectively, confirming its superior performance and reliability. Thus, the proposed method can provide a trustworthy fault diagnosis tool for mechanical systems in industrial scenarios considering anomalous sensor data.

Details DOI

AAAI Conference 2024 Conference Paper

Deep Active Learning with Noise Stability

Xingjian Li
Pengkun Yang
Yangcheng Gu
Xueying Zhan
Tianyang Wang
Min Xu
Chengzhong Xu

Uncertainty estimation for unlabeled data is crucial to active learning. With a deep neural network employed as the backbone model, the data selection process is highly challenging due to the potential over-confidence of the model inference. Existing methods resort to special learning fashions (e.g. adversarial) or auxiliary models to address this challenge. This tends to result in complex and inefficient pipelines, which would render the methods impractical. In this work, we propose a novel algorithm that leverages noise stability to estimate data uncertainty. The key idea is to measure the output derivation from the original observation when the model parameters are randomly perturbed by noise. We provide theoretical analyses by leveraging the small Gaussian noise theory and demonstrate that our method favors a subset with large and diverse gradients. Our method is generally applicable in various tasks, including computer vision, natural language processing, and structural data analysis. It achieves competitive performance compared against state-of-the-art active learning baselines.

PDF Details DOI

AAAI Conference 2022 Conference Paper

Boosting Active Learning via Improving Test Performance

Tianyang Wang
Xingjian Li
Pengkun Yang
Guosheng Hu
Xiangrui Zeng
Siyu Huang
Cheng-Zhong Xu
Min Xu

Central to active learning (AL) is what data should be selected for annotation. Existing works attempt to select highly uncertain or informative data for annotation. Nevertheless, it remains unclear how selected data impacts the test performance of the task model used in AL. In this work, we explore such an impact by theoretically proving that selecting unlabeled data of higher gradient norm leads to a lower upper-bound of test loss, resulting in a better test performance. However, due to the lack of label information, directly computing gradient norm for unlabeled data is infeasible. To address this challenge, we propose two schemes, namely expected-gradnorm and entropy-gradnorm. The former computes the gradient norm by constructing an expected empirical loss while the latter constructs an unsupervised loss with entropy. Furthermore, we integrate the two schemes in a universal AL framework. We evaluate our method on classical image classification and semantic segmentation tasks. To demonstrate its competency in domain applications and its robustness to noise, we also validate our method on a cellular imaging analysis task, namely cryo-Electron Tomography subtomogram classification. Results demonstrate that our method achieves superior performance against the state of the art. We refer readers to https: //arxiv. org/pdf/2112. 05683. pdf for the full version of this paper which includes the appendix and source code link.

PDF Details