魏坤，副教授，硕士生导师，主要研究方向为增量学习、多模态学习及视觉内容生成。共发表论文30余篇，均为CCF-A类会议或SCI一区期刊，研究成果已被谷歌学术检索引用1200余次。

主持或参与多项国家及省部级科研项目，包括国家重点研发计划重点专项、国家自然科学基金重点/面上/青年项目、陕西省重点研发计划项目，以及西安电子科技大学杭研院概念基金和校级科研项目。担任TPAMI、TNNLS、TIP等十余种国际权威期刊审稿人，并担任ICML、NeurIPS、ICLR、CVPR、ICCV、AAAI等国际顶级会议的程序委员会委员。

谷歌学术：https://scholar.google.com/citations?hl=zh-CN&user=fju0aWQAAAAJ

主要合作者：邓成教授、武阿明教授、杨旭副教授、杨二昆副教授

研究成果：

Under Review:

Rotation Control Unlearning: Quantifying and Controlling Continuous Unlearning for LLM with The Cognitive Rotation Space, Xiang Zhang, Kun Wei, Xu Yang, Chenghao Xu, Su Yan, Cheng Deng, ICLR2026.

A Bridge from Audio to Video: Phoneme-Viseme Alignment Allows Every Face to Speak Multiple Languages,  Zibo Su, Kun Wei, J Li, X Yang, C Deng, ICLR2026.

Perceive, Reflect and Understand Long Video: Progressive Multi-Granular Clue Exploration with Interactive Agents, Jiahua Li, Kun Wei, Zhe Xu, Zibo Su, Xu Yang, Cheng Deng,  ICLR2026.

Trajectory Regularized: Reducing Storage Dependency in Model Merging for Continual Learning, Xi Wang, Xu Yang, Kun Wei, Cheng Deng, CVPR2026.

2026年:

Towards Robust Edge Model Adaptation via Elastic Architecture Search, Xianhang Chu, Xu Yang, Kun Wei, Xi Wang, AAAI2026.（oral）

Diffusion-calibrated Continual Test-time Adapation, Xu Yang, Moqi Li, Kun Wei, AAAI2026.（oral）

Backtrace Mamba: Reviving Critical Temporal Contexts via Hierarchical Memory Compression for Online Action Detection, Su Yan, Jiahua Li, Kun Wei, Cheng Deng, AAAI2026.

2025年

Robust Temporal Action Localization With Meta Boundary Refinement, Jiahua Li, Kun Wei, Zhe Xu, Liejun  Wang, Cheng Deng, IEEE Transactions on Multimedia.

Class Incremental Learning via Contrastive Complementary Augmentation, Xu Wang, Xu Yang, Kun Wei, Yanan Gu, Cheng Deng, IEEE Transactions on Image Processing.

Outstanding orthodontist: no more artifactual teeth in talking face, Zibo Su, Ziqi Zhang, Kun Wei*, Xu Yang, Cheng Deng, IJCAI2025.

Q-MiniSAM2: A Quantization-based Benchmark for Resource-Efficient Video Segmentation, Xuanxuan Ren, Xiangyu Li, Kun Wei, Xu Yang, Yanhua Yang, IJCAI2025.

Compress to One Point: Neural Collapse for Pre-Trained Model-Based Class-Incremental Learning, Kun Wei, Zhe Xu, Cheng Deng, AAAI2025.

Energy vs. Noise: Towards Robust Temporal Action Localization in Open-World, Chenyu Mu, Jiahua Li, Kun Wei*, Cheng Deng, AAAI2025.

Dynamic Adapter Tuning for Long-Tailed Class-Incremental Learning,Yanan Gu, Muli Yang, Xu Yang, Kun Wei, Hongyuan Zhu, Gabriel James Goenawan, Cheng Deng, WACV2025.

2024年

ACFun: Abstract-Concrete Fusion Facial Stylization, Jiapeng Ji, Kun Wei*, Ziqi Zhang, Cheng Deng, NeurIPS2024.

Exploiting Intrinsic Multilateral Logical Rules for Weakly Supervised Natural Language Video Localization, Zhe Xu, Kun Wei, Xu Yang, Cheng Deng, ACL2024.

Unsupervised Graph Transformer with Augmentation-Free Contrastive Learning, Han Zhao, Xu Yang, Kun Wei, Cheng Deng, Dacheng Tao, IEEE Transactions on Knowledge and Data Engineering.

Few-Shot Generative Model Adaptation via Optimal Kernel Modulation, Ziqi Zhang, Siduo Pan, Kun Wei, Jiapeng Ji, Xu Yang, Cheng Deng, IEEE Transactions on Circuits and Systems for Video Technology.

Few-Shot Generative Model Adaptation via Style-Guided Prompt,  Siduo Pan, Ziqi Zhang, Kun Wei*, Xu Yang, Cheng Deng, IEEE Transactions on Multimedia.

Class-Incremental Unsupervised Domain Adaptation via Pseudo-Label Distillation, Kun Wei, Xu Yang, Zhe Xu, Cheng Deng, IEEE Transactions on Image Processing.

Do You Guys Want to Dance: Zero-Shot Compositional Human Dance Generation with Multiple Persons, Zhe Xu, Kun Wei, Xu Yang, Cheng Deng, arXiv

Navigating Continual Test-Time Adaptation with Symbiosis Knowledge, Xu Yang, Moqi Li, Jie Yin, Kun Wei, Cheng Deng, IJCAI2024.

A Versatile Framework for Continual Test-Time Domain Adaptation: Balancing Discriminability and Generalizability, Xu Yang, Xuan Chen, Moqi Li, Kun Wei, Cheng Deng, CVPR2024.

Long-Tail Class Incremental Learning via Independent Sub-Prototype Construction, Xi Wang, Xu Yang, Jie Yin, Kun Wei, Cheng Deng, CVPR2024.

2023年

Robust Commonsense Reasoning Against Noisy Labels Using Adaptive Correction, TCYB

CSC-GCN: Contrastive Semantic Calibration for Graph Convolution Network, JII

Hierarchical Prompt Learning for Compositional Zero-Shot Recognition, IJCAI

Exploring Safety Supervision for Continual Test-Time Domain Adaptation, IJCAI

Bilateral Relation Distillation for Weakly Supervised Temporal Action Localization, TPAMI

ALR-GAN: Adaptive Layout Refinement for Text-to-Image Synthesis, TMM

2022年

Point-Supervised Video Temporal Grounding, TMM

Incremental Embedding Learning with Disentangled Representation Translation, TNNLS

HiSA: Hierarchically Semantic Associating for Video Temporal Grounding, TIP

A Survey on Multimodal-Guided Visual Content Synthesis, Neurocomputing

ME-GAN: Learning Panoptic Electrocardio Representations for Multi-View ECG Synthesis Conditioned on Heart Diseases, ICML

Learning Universal Adversarial Perturbation by Adversarial Example, AAAI

Not Just Selection, But Exploration: Online Class-Incremental Continual Learning via Dual View Consistency, CVPR

2021年

Learning Semantic Priors for Texture-Realistic Sketch-to-Image Synthesis, Neurocomputing

Incremental Zero-Shot Learning, TCYB

Incremental Embedding Learning via Zero-Shot Translation, AAAI

Class-Incremental Instance Segmentation via Multi-Teacher Networks, AAAI

Generalized Zero-Shot Learning via Disentangled Representation, AAAI

Nearest Neighbor Matching for Deep Clustering, CVPR

SelfSAGCN: Self-Supervised Semantic Alignment for Graph Convolution Network, CVPR

2020年

Lifelong Zero-Shot Learning, IJCAI

Adversarial Learning for Robust Deep Clustering, NeurIPS

2019年

Adversarial Fine-Grained Composition Learning for Unseen Attribute-Object Recognition, ICCV