Posts by Collection

portfolio

publications

FPD: Feature Pyramid Knowledge Distillation

Published in Neural Information Processing (ICONIP), Lecture Notes in Computer Science, 2023

Feature-pyramid knowledge distillation, published in the Neural Information Processing (ICONIP) proceedings, Lecture Notes in Computer Science (2023).

Recommended citation: Qi Wang, Lu Liu, Wenxin Yu, Zhiqiang Zhang, Yuxin Liu, Shiyu Cheng, Xuewen Zhang, and Jun Gong. (2023). "FPD: Feature Pyramid Knowledge Distillation." Neural Information Processing, Lecture Notes in Computer Science, 100-111.
Publisher Page

Image Inpainting with Semantic-Aware Transformer

Published in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023

Image inpainting with a semantic-aware Transformer, presented at ICASSP 2023.

Recommended citation: Shiyu Chen, Wenxin Yu, Qi Wang, Jun Gong, and Peng Chen. (2023). "Image Inpainting with Semantic-Aware Transformer." IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 1-5.
Publisher Page

BCKD: Block-Correlation Knowledge Distillation

Published in IEEE International Conference on Image Processing (ICIP), 2023

Block-correlation knowledge distillation, presented at IEEE ICIP 2023.

Recommended citation: Qi Wang, Lu Liu, Wenxin Yu, Shiyu Chen, Jun Gong, and Peng Chen. (2023). "BCKD: Block-Correlation Knowledge Distillation." IEEE International Conference on Image Processing (ICIP), 3225-3229.
Publisher Page

TSFC: Texture and Structure Features Coupling for Image Inpainting

Published in IEEE International Conference on Image Processing (ICIP), 2023

Image inpainting by coupling texture and structure features, presented at IEEE ICIP 2023.

Recommended citation: Lu Liu, Qi Wang, Wenxin Yu, Shiyu Chen, Jun Gong, and Peng Chen. (2023). "TSFC: Texture and Structure Features Coupling for Image Inpainting." IEEE International Conference on Image Processing (ICIP), 3279-3283.
Publisher Page

Similarity Knowledge Distillation with Calibrated Mask

Published in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024

Similarity-based knowledge distillation with a calibrated mask, presented as a poster at ICASSP 2024.

Recommended citation: Qi Wang, Wenxin Yu, Lu Che, Chang Liu, Zhiqiang Zhang, Jun Gong, and Peng Chen. (2024). "Similarity Knowledge Distillation with Calibrated Mask." IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 5770-5774.
Publisher Page

Axial Attention Transformer for Fast High-quality Image Style Transfer

Published in IEEE International Symposium on Circuits and Systems (ISCAS), 2024

An axial-attention Transformer for fast, high-quality image style transfer, presented at IEEE ISCAS 2024.

Recommended citation: Yuxin Liu, Wenxin Yu, Zhiqiang Zhang, Qi Wang, and Lu Che. (2024). "Axial Attention Transformer for Fast High-quality Image Style Transfer." IEEE International Symposium on Circuits and Systems (ISCAS), 1-5.
Publisher Page

Top-KD: Top-scaled contrastive knowledge distillation

Published in Neurocomputing, 2026

A contrastive knowledge distillation method published in Neurocomputing (vol. 685, article 133636).

Recommended citation: Qi Wang and Jinjia Zhou. (2026). "Top-KD: Top-scaled contrastive knowledge distillation." Neurocomputing 685, 133636.
Publisher Page

Training-Free Adaptive Sampling Scheduler for Discrete-Time Consistency Distillation

Published in International Conference on Artificial Neural Networks (ICANN), 2026

Few-step consistency models fail not because the student is weak but because the sampling schedule is rigid. We score every timestep by an SNR-derived Importance measure, spend the same step budget where the trajectory actually moves, and add a matching sampler and colour-balancing pass. No retraining is required, and the method drops into LCM, PCM, TCD and TDD. At 2 sampling steps FID falls by up to 86%.

Recommended citation: Qi Wang and Jinjia Zhou. (2026). "Training-Free Adaptive Sampling Scheduler for Discrete-Time Consistency Distillation." International Conference on Artificial Neural Networks (ICANN).
Publisher Page | Download Bibtex

talks

teaching