Author name cluster

Fei Du

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.

8 papers

1 author row

AAMAS Conference 2026 Conference Paper

CQGen-MAS: A Multi-Agent System for Competency Questions Generation from Ontology

Fei Du
Huilong An
Feng Li
Li Chen
Weidong Liu

Competency questions (CQs) delineate an ontology’s intended scope and use, yet many published ontologies are released without CQs, hindering interpretation and reuse. We present CQGen-MAS, a multi-agent system that automatically generates CQs from an existing ontology. We evaluate CQGen-MAS on a diverse set of ontologies and compare it with baseline approaches. Experimental results show that CQGen-MAS consistently performs favorably across diverse ontologies, indicating improved coverage and stronger semantic alignment with their conceptual scope.

PDF

AAAI Conference 2025 Conference Paper

RealisID: Scale-Robust and Fine-Controllable Identity Customization via Local and Global Complementation

Zhaoyang Sun
Fei Du
Weihua Chen
Fan Wang
Yaxiong Chen
Yi Rong
Shengwu Xiong

Recently, the success of text-to-image synthesis has greatly advanced the development of identity customization techniques, whose main goal is to produce realistic identity-specific photographs based on text prompts and reference face images. However, it is difficult for existing identity customization methods to simultaneously meet the various requirements of different real-world applications, including the identity fidelity of small face, the control of face location, pose and expression, as well as the customization of multiple persons. To this end, we propose a scale-robust and fine-controllable method, namely RealisID, which learns different control capabilities through the cooperation between a pair of local and global branches. Specifically, by using cropping and up-sampling operations to filter out face-irrelevant information, the local branch concentrates the fine control of facial details and the scale-robust identity fidelity within the face region. Meanwhile, the global branch manages the overall harmony of the entire image. It also controls the face location by taking the location guidance as input. As a result, RealisID can benefit from the complementarity of these two branches. Finally, by implementing our branches with two different variants of ControlNet, our method can be easily extended to handle multi-person customization, even only trained on single-person datasets. Extensive experiments and ablation studies indicate the effectiveness of RealisID and verify its ability in fulfilling all the requirements mentioned above.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

SHMT: Self-supervised Hierarchical Makeup Transfer via Latent Diffusion Models

Zhaoyang Sun
Shengwu Xiong
Yaxiong Chen
Fei Du
Weihua Chen
Fan Wang
Yi Rong

This paper studies the challenging task of makeup transfer, which aims to apply diverse makeup styles precisely and naturally to a given facial image. Due to the absence of paired data, current methods typically synthesize sub-optimal pseudo ground truths to guide the model training, resulting in low makeup fidelity. Additionally, different makeup styles generally have varying effects on the person face, but existing methods struggle to deal with this diversity. To address these issues, we propose a novel Self-supervised Hierarchical Makeup Transfer (SHMT) method via latent diffusion models. Following a "decoupling-and-reconstruction" paradigm, SHMT works in a self-supervised manner, freeing itself from the misguidance of imprecise pseudo-paired data. Furthermore, to accommodate a variety of makeup styles, hierarchical texture details are decomposed via a Laplacian pyramid and selectively introduced to the content representation. Finally, we design a novel Iterative Dual Alignment (IDA) module that dynamically adjusts the injection condition of the diffusion model, allowing the alignment errors caused by the domain gap between content and makeup representations to be corrected. Extensive quantitative and qualitative analyses demonstrate the effectiveness of our method. Our code is available at https: //github. com/Snowfallingplum/SHMT.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

Data Pruning via Moving-one-Sample-out

Haoru Tan
Sitong Wu
Fei Du
Yukang Chen
Zhibin Wang
Fan Wang
Xiaojuan Qi

In this paper, we propose a novel data-pruning approach called moving-one-sample-out (MoSo), which aims to identify and remove the least informative samples from the training set. The core insight behind MoSo is to determine the importance of each sample by assessing its impact on the optimal empirical risk. This is achieved by measuring the extent to which the empirical risk changes when a particular sample is excluded from the training set. Instead of using the computationally expensive leaving-one-out-retraining procedure, we propose an efficient first-order approximator that only requires gradient information from different training stages. The key idea behind our approximation is that samples with gradients that are consistently aligned with the average gradient of the training set are more informative and should receive higher scores, which could be intuitively understood as follows: if the gradient from a specific sample is consistent with the average gradient vector, it implies that optimizing the network using the sample will yield a similar effect on all remaining samples. Experimental results demonstrate that MoSo effectively mitigates severe performance degradation at high pruning ratios and achieves satisfactory performance across various settings. Experimental results demonstrate that MoSo effectively mitigates severe performance degradation at high pruning ratios and outperforms state-of-the-art methods by a large margin across various settings.

PDF Details

AAAI Conference 2023 Conference Paper

SwinRDM: Integrate SwinRNN with Diffusion Model towards High-Resolution and High-Quality Weather Forecasting

Lei Chen
Fei Du
Yuan Hu
Zhibin Wang
Fan Wang

Data-driven medium-range weather forecasting has attracted much attention in recent years. However, the forecasting accuracy at high resolution is unsatisfactory currently. Pursuing high-resolution and high-quality weather forecasting, we develop a data-driven model SwinRDM which integrates an improved version of SwinRNN with a diffusion model. SwinRDM performs predictions at 0.25-degree resolution and achieves superior forecasting accuracy to IFS (Integrated Forecast System), the state-of-the-art operational NWP model, on representative atmospheric variables including 500 hPa geopotential (Z500), 850 hPa temperature (T850), 2-m temperature (T2M), and total precipitation (TP), at lead times of up to 5 days. We propose to leverage a two-step strategy to achieve high-resolution predictions at 0.25-degree considering the trade-off between computation memory and forecasting accuracy. Recurrent predictions for future atmospheric fields are firstly performed at 1.40625-degree resolution, and then a diffusion-based super-resolution model is leveraged to recover the high spatial resolution and finer-scale atmospheric details. SwinRDM pushes forward the performance and potential of data-driven models for a large margin towards operational applications.

PDF Details DOI

YNICL Journal 2020 Journal Article

The structural basis for interhemispheric functional connectivity: Evidence from individuals with agenesis of the corpus callosum

Junliang Yuan
Xiaopeng Song
Elliot Kuan
Shuangkun Wang
Long Zuo
Dost Ongur
Wenli Hu
Fei Du

Agenesis of the corpus callosum (AgCC) is a rare congenital malformation characterized by partial or complete absence of the corpus callosum (CC). The effects of AgCC on cerebral structural and functional networks are not clear. We aimed to utilize AgCC as a model to characterize the relationship between brain structure and function. Diffusion tensor imaging and resting-state fMRI data were collected from nine AgCC and ten healthy subjects. The interhemispheric functional connectivity (FC) was quantified using a voxel-mirrored-homotopic-connectivity (VMHC) method, and its correlation with the number (FN) and fractional anisotropy (FA) of the fibers crossing the CC was calculated. Graph-based network analyses of structural and functional topologic properties were performed. AgCC subjects showed markedly reduced VMHC compared to controls. VMHC was significantly correlated with the FN and FA of the fibers crossing the CC. Structural network analyses revealed impaired global properties, but intact local properties in AgCC compared to controls. Functional network analyses showed no significant difference in network properties between the groups. Finally, in both groups, brain areas with more fiber connections were more likely to build a positive FC with each other, while areas with decreased white matter connections were more likely to result in negative FC. Our observations demonstrate that interhemispheric FC is highly dependent on CC structure. Increased alternative intrahemispheric SC might be a compensatory mechanism in AgCC that helps to maintain normal global brain function. Our study provides insights into the underlying neurological pathophysiology of brain malformations, thereby helping to elucidate the structure-function relationship of normal human brain.

Details DOI

YNIMG Journal 2015 Journal Article

Frontal P3 event-related potential is related to brain glutamine/glutamate ratio measured in vivo

Mei-Hua Hall
J. Eric Jensen
Fei Du
Jordan W. Smoller
Lauren O’Connor
Kevin M. Spencer
Dost Öngür

Background The auditory P3 event-related potential (ERP) is thought to index cognitive processing relevant to attention and working memory processes. Drug challenge studies suggest that glutamate neurotransmission plays an important role in modulating P3 ERP. However, while direct links between glutamate activity and P3 ERP response in humans are suspected, mechanistic details remain largely unknown. We investigated here the relationships between P3 ERP and indices of glutamatergic processing measured in vivo with proton magnetic resonance spectroscopy (1H MRS). We hypothesized that a higher index of glutamatergic processing (glutamine/glutamate ratio; abbreviated Gln/Glu) in the anterior cingulate (ACC) and in the parietal–occipital (POC) cortices would associate with larger frontal P3a and parietal P3b amplitudes, respectively. Methods Frontal P3a (Fz) and parietal P3b (Pz) were collected from 32 healthy participants who performed an auditory oddball task. Resting glutamate (Glu), glutamine (Gln), and Gln/Glu (an index of glutamatergic processing) measures were obtained on a 4T MR scanner using J-resolved MR spectroscopy. Linear regression and partial correlations were used for statistical analysis. Results Significant positive correlations were found between frontal P3a amplitude and ACC Gln/Glu ratio (partial R =0. 57; P =0. 001) and between frontal P3a amplitude and ACC Gln concentration (partial R =0. 43; P =0. 02). Relationships between parietal P3b and the glutamate indices in the POC were not significant. Conclusions These results indicate a specific connection between an index of glutamate neurotransmitter function in ACC and frontal P3 ERP, providing a novel insight into the neurochemistry underlying scalp recorded EEG response. Abnormalities in glutamate neurotransmission have been observed in schizophrenia and other psychiatric conditions and may underlie illness related deficits of P3 ERP.

Details DOI

YNIMG Journal 2012 Journal Article

Quantitative imaging of energy expenditure in human brain

Xiao-Hong Zhu
Hongyan Qiao
Fei Du
Qiang Xiong
Xiao Liu
Xiaoliang Zhang
Kamil Ugurbil
Wei Chen

Despite the essential role of the brain energy generated from ATP hydrolysis in supporting cortical neuronal activity and brain function, it is challenging to noninvasively image and directly quantify the energy expenditure in the human brain. In this study, we applied an advanced in vivo 31P MRS imaging approach to obtain regional cerebral metabolic rates of high-energy phosphate reactions catalyzed by ATPase (CMRATPase) and creatine kinase (CMRCK), and to determine CMRATPase and CMRCK in pure gray mater (GM) and white mater (WM), respectively. It was found that both ATPase and CK rates are three times higher in GM than WM; and CMRCK is seven times higher than CMRATPase in GM and WM. Among the total brain ATP consumption in the human cortical GM and WM, 77% of them are used by GM in which approximately 96% is by neurons. A single cortical neuron utilizes approximately 4. 7billion ATPs per second in a resting human brain. This study demonstrates the unique utility of in vivo 31P MRS imaging modality for direct imaging of brain energy generated from ATP hydrolysis, and provides new insights into the human brain energetics and its role in supporting neuronal activity and brain function.

Details DOI