WEI Tianwen 魏天闻

Contact Information:

weitianwen AT xiaomi.com

Background

• 2002-2006: BSc in applied mathematics, University of Wuhan, Wuhan, China.

• 2007-2009: MSc in applied mathematics, Université Lille 1, Lille, France.

• 2009-2013: PhD in applied mathematics, Université Lille 1, Lille, France.

• 2014-2015: Postdoctoral researcher at Université Franche-Comté, Besançon, France.

• 2014-2018: Assistant Professor at Zhongnan University of Economics and Law, Wuhan, China.

• 2018-present: Senior NLP Researcher at Xiaomi Inc.

Research Interests:

• Natural Language Processing

• Machine learning

• Independent component analysis

• Asymptotic statistics

Publications:

• T. Wei and J. Qi, A flexible multi-task model for BERT serving, ACL 2022.

• T. Wei and J. Qi and S. He, Masked Conditional Random Fields for sequence labeling, NAACL 2021.

• T. Wei and S. Chretien, A penalized autoencoder approach for nonlinear independent component analysis, ICASSP 2019.

• S. Chretien and T. Wei, "Sensing tensors with Gaussian filters", IEEE transaction on Information Theory, pp. 843-852, vol. 63, Issue 2, Feb.2017, DOI 10.1109/TIT.2016.2633413.

• S. Chretien and T. Wei, "The subdifferential of some tensor norms", Linear Algebra and its Applications, DOI 10.1016/j.laa.2017.02.003. [PDF]

• S. Chretien and T. Wei, "Convex recovery of tensors using nuclear norm penalization", the 12th International Conference on Latent Variable Analysis and Source Separation (LVA/ICA 2015), Liberec, Czech Republic. [ArXiv]

• T. Wei, "An overview of the asymptotic performance of the family of the FastICA algorithms", the 12th International Conference on Latent Variable Analysis and Source Separation (LVA/ICA 2015), Liberec, Czech Republic. [ArXiv]

• S. Chretien and T. Wei, "Von Neumann's trace inequality for tensors", Linear algebra and its applications, 2015. [ArXiv]

• T. Wei, "A convergence and asymptotic analysis of the generalized FastICA algorithm". IEEE transaction on Signal Processing. [ArXiv]

• T. Wei, "A study of the fixed points and spurious solutions of the FastICA algorithm", Neural Computing and Applications. [ArXiv]

• T. Wei, “Asymptotic analysis of the generalized FastICA algorithm”, 2014 IEEE workshop on statistical signal processing, Gold Coast, Australia. [PDF]

• T. Wei, “On the spurious solutions of the FastICA algorithm,” 2014 IEEE workshop on statistical signal processing, Gold Coast, Australia. [PDF]

• A. Dermoune and T. Wei, “FastICA algorithm: Five criteria for the optimal choice of the nonlinearity function,” IEEE Transactions on Signal Processing, vol. 61, no. 8, pp. 2078–2087, Apr. 2013. [PDF]

• A. Dermoune, N. Rahmania, and T. Wei, “General linear mixed model and signal extraction problem with constraint,” Journal of Multivariate Analysis, vol. 105, pp. 311–321, 2012. [PDF]