Wei Huang

Wei Huang

Auto-generated profile, based on publicly available information

Wei Huang

School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
7 February 2023 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Verified
Journal Article
Closed access
ZPZhiliang Peng
Zhiliang Peng
ZGZonghao Guo
Zonghao Guo
WHWei Huang
Wei Huang
YWYaowei Wang
Yaowei Wang
LXLingxi Xie
Lingxi Xie
QYQixiang Ye
Qixiang Ye
With convolution operations, Convolutional Neural Networks (CNNs) are good at extracting local features but experience difficulty to capture global representations. With cascaded self-attention modules, vision transformers can capture long-distance feature dependencies but unfortunately deteriorate local feature details. In this paper, we propose a hybrid network structure, termed Conformer, to take both advantages of convolution operations and self-attention mechanisms for enhanced representation learning. Conformer roots in feature coupling of CNN local features and transformer global representations under different resolutions in an interactive fashion. Conformer adopts a dual structure so that local details and global dependencies are retained to the maximum extent. We also propose a Conformer-based detector (ConformerDet), which learns to predict and refine object proposals, by performing region-level feature coupling in an augmented cross-attention fashion. Experiments on ImageNet and MS COCO datasets validate Conformer's superiority for visual recognition and object detection, demonstrating its potential to be a general backbone network.

25 April 2024 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
WHWei Huang
Wei Huang
ZPZhiliang Peng
Zhiliang Peng
LDLi Dong
Li Dong
FWFuru Wei
Furu Wei
QYQixiang Ye
Qixiang Ye
JJJianbin Jiao
Jianbin Jiao
To transfer the representation capacity of large pre-trained models to lightweight models, knowledge distillation has been widely explored. However, conventional single-stage distillation methods are prone to getting stuck in the transfer of task-specific knowledge, making it difficult to retain task-agnostic knowledge which is crucial for model generalization. In this study, we propose generic-to-specific distillation (G2SD), to boost lightweight models under the assistance of large models pre-trained by masked image modeling. In generic distillation, the decoder of a small model is encouraged to align feature predictions with that of a large model, so that task-agnostic knowledge can be transferred. In specific distillation, predictions of the small model are encouraged to be consistent with those of the large model, to guarantee task performance. G2SD is also applicable for heterogeneous settings( i.e ., distilling from ViT to CNN). With G2SD, the ViT-Small model respectively achieves 98.9%, 98.4%, 99.3% and 98.9% accuracies when compared with its teachers (ViT-Base) for image classification, object detection, semantic segmentation and video recognition tasks. The lightweight ResNet models are improved to a new height on image classification task. The code is available at github.com/pengzhiliang/G2SD.

13 August 2021
Verified
Conference Paper
Open access
YCyuwen chen
yuwen chen
YLYuJie Li
YuJie Li
WHWei Huang
Wei Huang
JZJu Zhang
Ju Zhang
Bin YiBin Yi
Bin Yi
XQXiaolin Qin
Xiaolin Qin
The occurrence of perioperative critical adverse events will affect the quality of medical services and threaten the safety of patients. Using scientific methods to assess the risk of critical illness in perioperative period is of great significance to improve the quality of medical service and ensure the safety of patients. However, the diagnosis and treatment data of perioperative patients are multi-source and irregular, and only one physiological information can not accurately reflect the patient's condition. In previous studies, it is found that a variety of physiological information can transmit the information of human health or not, which can be used to evaluate critical illness and physical condition. Therefore, this paper integrates the preoperative clinical structure data, intraoperative vital signs monitoring time series data and intraoperative anesthesia event time series data. Based on deep learning technology, the multi-modal data of patients are embedded and mapped into the same recessive semantic space to realize the real-time tracking and early warning of severe events, reduce postoperative complications and improve the early diagnosis efficiency of critical adverse events. The results showed that the performance of the model based on the multi-modal data was better than that based on the real military data.