My name is Yu-Chia Chen, and I am a Research Scientist at Meta working on GenAI for Ads. My research interests are in unsupervised learning, geometric data analysis, and vector field learning. Please check out my CV here.
Education
I graduated with my Ph.D. in Electrical Engineering at the University of Washington (UW) in August 2021. I was honored to complete my thesis, Learning Topological Structures and Vector Fields on Manifolds with (Higher-order) Discrete Laplacians, under the supervision of professor Marina Meilă.
I obtained my B.S. in Physics at National Taiwan University (NTU) in June 2015. I was pleased to work with professor Yang-Fang Chen on the bio-inspired random laser.
Experience
I was at Facebook as a Machine Learning Intern in 2020. We built transfer learning and multi-task learning deep learning models to optimize the click-through rate (CTR) based recommendation system for search ads placement.
I interned at Microsoft Research during the summer of 2018. We developed modeling large-scale temporal networks by dynamic stochastic block model and its extension to causal impact on dynamic social networks. Our work has been published in KDD’19 and Animal Behaviour.
News
Sep 2021 Paper The decomposition of the higher-order homology embedding constructed from the k-Laplacian accepted as an oral presentation at NeurIPS’21 (acceptance rate 1%)
Sep 2021 New paper Social connectedness and movements among communities of giraffes vary by sex and age class accepted to Animal Behaviour
- Jul 2021 New paper on arXiv: The decomposition of the higher-order homology embedding constructed from the k-Laplacian
- Jun 2021 New paper on ChemRxiv: Water-Accelerated Photo-oxidation of CH3NH3PbI3 Perovskite: Mechanism, rate orders, and rate constants
- Mar 2021 New paper on arXiv: Helmholtzian Eigenmap: Topological feature discovery & edge flow learning from point cloud data
- Jun 2020 Machine learning internship at Facebook, Seattle, WA
- Dec 2019 Presenting at NeurIPS’19, Vancouver, Canada (12/08/19 – 12/14/19)
- Nov 2019 Workshop paper Manifold coordinates with physical meaning accepted for poster presentation at the Machine Learning and the Physical Sciences (ML4PS) Workshop at NeurIPS’19
- Sep 2019 Paper Selecting the independent coordinates of manifolds with large aspect ratios accepted for poster presentation at NeurIPS’19 (acceptance rate 21.2%)
- Sep 2019 Attending the IPAM long program: Machine Learning for Physics and the Physics of Learning at UCLA (09/04/19 – 12/08/19)
- Aug 2019 Presenting at KDD’19, Anchorage, AK (08/04/19 – 08/08/19)
- Jun 2019 New paper on arXiv: Selecting the independent coordinates of manifolds with large aspect ratios
- Apr 2019 New paper On Dynamic Network Models and Application to Causal Impact accepted for poster presentation at the KDD’19 Research Track (acceptance rate 14.2%)