你可能对以下一个或多个主题感兴趣,部分主题可能涉及机器学习理论、黑盒优化等较为数学化的内容:
§ Hyperbolic Geometry (双曲空间和黎曼几何的建模与优化)
[1] Xiaofeng Cao, Ivor W. Tsang. Distribution Disagreement via Lorentzian Focal Representation, IEEE Transactions on Pattern Analysis and Machine Intelligence. (T-PAMI为模式分析旗舰期刊,IF=24.314)
[2] Xiaofeng Cao, Ivor W. Tsang. Hyperbolic Fréchet Mean. Pattern Recognition. IF=8.518. (独著工作,研究非欧空间下的Fréchet Mean优化与求解)
[3] Xiaofeng Xu, Ivor W. Tsang, Xiaofeng Cao, et al. Learning image-specific attributes by hyperbolic neighborhood graph propagation. International Joint Conference on Artificial Intelligence 2019.(非欧树结构)
[4] Yang Tao and Xiaofeng Cao. Perturbation Elimination via Homeomorphic Manifold Tubes, IEEE Transactions on Neural Networks and Learning Systems, review. (黎曼通管优化)
§ Machine Teaching/ Black-box/Bayesian Optimization (机器教学/黑盒/贝叶斯优化)
[1] Xiaofeng Cao, Ivor W. Tsang. Distribution-based Machine Teaching for a Black-box, IEEE Transactions on Neural Networks and Learning Systems, IF=14.255.(黑盒机器教学优化理论)
[2] Xiaofeng Cao, Ivor W. Tsang. On the Geometry of Deep Bayesian Active Learning, IEEE Transactions on Emerging Topics in Computational Intelligence, revision, IF=4.851.
[3] Chen Zhang, Xiaofeng Cao. Pseudo-Iterative Machine Teaching. Pseudo-Iterative Machine Teaching, Artificial Intelligence (人工智能旗舰期刊,国际人工智能会刊), review. IF=14.05.(封闭梯度优化)
[4] Xiaofeng Cao# and Yaming Guo#. Black-box Teaching an Active Learner, Journal of Machine Learning Research, revision (机器学习领域旗舰期刊). 这可能是黑盒泛化理论重要的研究成果!
[5]Chen Zhang, Xiaofeng Cao. Functional optimization of machine teaching.(泛函梯度优化理论)
§ Active Learning Theory and Its Generalization Analysis (主动学习理论与其泛化分析)
[1] Xiaofeng Cao, Ivor W. Tsang. Shattering distribution for active learning, IEEE Transactions on Neural Networks and Learning Systems, 2020, IF=14.255. (分布破碎)
[2] Xiaofeng Cao, Ivor W. Tsang, Jianliang Xu. Cold-start Active Sampling via $\gamma$-Tube, IEEE Transactions on Cybernetics, 2021. IF=19.118.(通管采样)
§ Learning on Small Data (新主题:Small data is the future of AI)
[1] Xiaofeng Cao, Ivor W. Tsang. Learning on Small data via Minimizing Hyperspherical Energy.IEEE Transactions on Pattern Analysis and Machine Intelligence, revision.(最小能量球优化问题)
[2] Learning on Small Data: Transfer the Future of Artificial Intelligence to Now, Survey work, IEEE Transactions on Pattern Analysis and Machine Intelligence, review. (小数据学习)
§ General AI Applications: Data, Image, Graph.
[1] Xiaofeng Cao et al. Multidimensional Balance-Based Cluster Boundary Detection for High-Dimensional Data, IEEE Transactions on Neural Networks and Learning Systems, 30(6): 1867-1880, 2019. IF=14.255. (边界检测)
[2] Zenglin Shi, Le Zhang, Yun Liu, Xiaofeng Cao et al. Crowd counting with deep negative correlation learning. Computer Vision and Pattern Recognition Conference (CVPR) 2018.
[3]Yu Wang, Liang Hu, Wanfu Gao*, Xiaofeng Cao*, Yi Chang, “AdaNS: Adaptive Negative Sampling for Unsupervised Graph Representation Learning”, Pattern Recognition, 2022.IF=8.518. (图表征)
注:如果你有意与我一同工作,并且拟攻读硕士研究生,你可能会直接参与以上课题。如果你有意攻读博士,我们将一起探索更为基础和前沿的机器学习内容,可能包括Meta-Learning、Distribution Optimization, 等。你可能与我、Ivor Tsang一起工作。Ivor Tsang是国际知名机器学者,他是NeurIPS 2021 Exp Chair, ACML 2021 Co-Chair, ICML 2021 Senior Area Chair,同时也是JMLR、MLJ、T-PAMI、JAIR等机器学习和人工智能旗舰期刊的Editor/Associate Editor。课题组已经与剑桥大学、斯坦福大学的一流同行建立合作,欢迎有志青年加入我们!