“As for the future, your task is not to foresee it, but to enable it.”
- Antoine de Saint Exupery (1900-1944)
Artificial Intelligence
AI Enablers for Ultra-Efficient Learning and Security
We designed a zero-shot learning algorithm to alleviate the requirement for large-amount of training data, published results in NeurIPS 2019.
NeurIPS is a top tier academic conference in the AI domain (according to Google Scholar, NeurIPS is ranked #1 in AI conferences/journals).
Based on this research, we lead main AI players in China and established a national industrial standard on deep learning with small data set (T/CESA 1034–2019).
J. Ni, S. Zhang, and H. Xie, “Dual adversarial semantics-consistent network for generalized zero-shot learning,” in Neural Information Processing Systems (NeurIPS), vol. 32, Dec. 2019, pp. 6143–6154.
We also designed numerous enabling technologies for AI to learn fast and efficiently. For instance, Mercury speeds up model learning by 218%, Guided Dropout significantly improves false positives, precision-recall curve and average precision without increasing the amount of computation.
Mercury: Fast and Optimal Device Placement for Large Deep Learning Models. H Xu, P Zhou, H Xie, Y Liao. Proceedings of the 52nd International Conference on Parallel Processing, pp. 412-422. 2023
Guided Dropout: Improving Deep Networks Without Increased Computation. Y Liu, Y Li, Z Xu, X Liu, H Xie, H Zeng. INTELLIGENT AUTOMATION AND SOFT COMPUTING 36 (3), 2519-2528
Celeritas: Fast optimizer for large dataflow graphs. H Xu, Y Liao, H Xie, P Zhou. arXiv preprint arXiv:2208.00184
Towards Secure Multi-Agent Deep Reinforcement Learning: Adversarial Attacks and Countermeasures. C Zheng, C Zhen, H Xie, S Yang. 2022 IEEE Conference on Dependable and Secure Computing (DSC), 1-8
AI for Understanding the Physical World
We study how Artificial Intelligence can help solve challenges in health care, by collaborating with distinguished researchers from top tier hospitals (AnDing Hospital, TianTan Hospital) and medical research institutions (CCMU).
On exploring multiplicity of primitives and attributes for texture recognition in the wild. W Zhai, Y Cao, J Zhang, H Xie, D Tao, ZJ Zha. in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 1, pp. 403-420, Jan. 2024
W. Zhai, Y. Cao, H. Xie and Z. -J. Zha, "Deep Texton-Coherence Network for Camouflaged Object Detection," in IEEE Transactions on Multimedia, vol. 25, pp. 5155-5165, 2023
TKN: Transformer-based Keypoint Prediction Network For Real-time Video Prediction. H Li, P Zhou, Y Lin, Y Hao, H Xie, Y Liao. arXiv preprint arXiv:2303.09807
W. Zhai, Y. Cao, Z.-J. Zha, H. Xie, and F. Wu, “Deep structure-revealed network for texture recognition,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2020, pp. 11 007–11 016.