林澤佑 Tse-Yu Lin
政治大學應用數學博士，目前服務於臺灣大學資料科學學位學程，講授Python程式設計課程。研究興趣為幾何重建、深度學習、HodgeRank 的應用。
副職是不學無術者，一開始攻讀投資組合最佳化問題，接著渡過了記憶體設計、奈米尺度材料力學及智慧理財新創公司等業界生活，在"被"投入到教學神經網路的講師狀態後，開始研究微分幾何與神經網路之間的跨領域應用。
Taipei, Taiwan
Education
PhD in Mathematical Sciences, National Chengchi University
September 2012 - January 2021
Dissertation: Manifold Embedding with Deep ResNet and A Study of the Continuity of HodgeRank
I have two main research topics:
a) HodgeRank and its real-world applications;
b) Manifold reconstruction using Neural Network.
For the first topic, I mainly deal with HodgeRank method and its generalization. A sufficient condition of continuity of HodgeRank was studied.
For the second topic, I reformulate a manifold interpolation algorithm as a learning process of deep residual network (ResNet) using the characterization property of affine orthogonal projections.
Visiting Scholar, Department of Mathematics, University of California San Diego
June 2019 - October2019
Studied on manifold interpolation/reconstruction based on composition of local affine projections
MS of Science in Mathematical Science, National Chengchi University
September 2010 - July 2012
Thesis: Stochastic Portfolio Optimization Models for the Stable Growth Benchmark Tracking
I focused on the combination between portfolio optimization problem and scenario trees. Moment matching method was introduced to provide a user-defined scenario tree which fits the first few moments of past data.
BS of Science in Mathematical Science, National Chengchi University
September 2006 - June 2010
Major in Judo (2nd Dan), double major in Karate (brown belt), minor in Slavic languages and applied mathematics. Besides, I am one of the co-founders of speciality coffee club in NCCU.
- Mathematical Finance
- Slavic Languages And Literatures
Experience
Adjunct Instructor, Department of Financial Engineering and Actuarial Mathematics, Soochow University
August 2020 - July 2021
- 微積分 Calculus
- 線性代數 Linear Algebra
- 數值分析 Numerical Analysis
Intern, Institute for Information Industry
September 2018 - February2019
- Integrated different framework of neural network model.
- Developed visualization scheme about the quality of medical data.
Research Assistant, NCTS USRP 2018
計算代數幾何專題 NCTS website
- Assisted four students from NTU, NTHU, and NCTU to join work in the topic of computational algebraic geometry.
- Implemented a classical-to-tropical transformer for objects in both classical algebraic geometry and tropical geometry. e.g., tropicalization of polynomial, tropical curve, and Newton polytope.
- Implemented an algorithm for the classification of A-D-E singularities (du Val singularities).
Research Assistant, NCTS USRP 2017
手機中的資料科學 NCTS website
- Assisted four students from different universities to join work in the field of real world data science issue.
- Implemented a smartphone application w/ Matlab and improve it into real-time application.
- Summarized some algorithms in medical image analysis from Kaggle Data Science Bowl 2017.
Research Engineer, Masdar Institute
August 2014 - May 2015
- Developed statistical analysis procedure of data from atomic force microscope (AFM).
Mathematical Consultant, M31 Technology
July 2012 – June 2014
- Studied papers on the topic of circuit dynamic and static optimization.
- Developed corresponding algorithms on VLSI circuit tuning via optimization method.
Teaching Assistant, NCCU
September 2010 - June 2013
September 2014 - June 2020
- 微積分 Calculus
- 線性代數 Linear Algebra
- 高等微積分 Advanced Calculus
- 實分析 Real Analysis
- 微分幾何 Differential Geometry
- 作業研究Operations Research
- 投資學 Investment Science
- 期貨與選擇權 Futures and Options
- 財務數學 Financial Mathematics
Projects
HodgeRank
HodgeRank is a pairwise comparison method for recommendation system.
The theory of HodgeRank is based on the theory of cohomology group and Hodge/Helmholtz Decomposition.
The figure is referred from the paper by Jiang, X., Lim, L. H., Yao, Y., & Ye, Y.
Medical Image Analysis
I am recently focus on how to use neural network in medical image analysis. I am working with master students in NTU with Prof. Weichung Wang. We are now trying to duplicate some existence models and then improve it in any sense.
This figure is the MRI image of my left shoulder.
Talks
- Manifold Reconstruction using Deep Residual Network
- 初探 Python AI 深度學習的第一堂課
- 深度學習理論與實務工作坊
- Index Tracking Models with Forecasting (with Ming Long Liu)
- Men and Women Matching Models and Its Applications (with Ming Long Liuand Po-Hsiang Chan)
- A brief introduction to HodgeRank
- 與學生互動的教學技巧
- Feature Portfolio Construction with Deep Learning (join work with Yen-lung Tsai)
- Advanced Keras Modeling Techniques
Publication
- Sun, H. C., Lin, T. Y., & Tsai, Y. L. (2023). Performance prediction in major league baseball by long short-term memory networks. International Journal of Data Science and Analytics, 15(1), 93-104.
- Sun, H. C., Lin, T. Y., & Tsai, Y. L. (2021). LSTM-based Approaches for the Performance Prediction in MLB, International Workshop on Domain-Driven Data Mining (DDDM). In conjunction with 21th SIAM International Conference on Data Mining (SDM21).
- Liu, H. K., Lin, T. Y., & Tsai, Y. L. (2020). On the Pricing Formula for the Perpetual American Volatility Option Under the Mean-reverting Processes. Taiwanese Journal of Mathematics. (SCI)
- Lin, T. Y., & Tsai, Y. L. Sun, H. C. (2020). Manifold Reconstruction with ResNet-Like Architecture. SIAM Conference on
Mathematics of Data Science (MDS20).
- Lin, T. Y., & Tsai, Y. L. (2019). Graph Laplacian Problems on Graph Neural Networks. The 19th SIAM International Conference on Data Mining (SDM19) Workshop on Deep Learning for Graphs.
- Lin, T. Y., & Tsai, Y. L. (2018). An Application of HodgeRank to the Online Peer Assessment, 4th International Workshop on Machine Learning Methods for Recommender Systems. In conjunction with 18th SIAM International Conference on Data Mining (SDM18).
- Lin, T. Y., & Tsai, Y. L. (2017). Invariant Generative Adversarial Network on Chinese characters. Doctoral consortium session at the 2017 conference on technologies and applications of artificial intelligence (TAAI 2017).
- Lin, T. Y., & Tsai, Y. L. (2017). Feature Portfolio Construction with Deep Learning, TWSIAM Annual Meeting
- Tsai, Y.L. & Lin, T.Y. (2016). Toward Deep Learning for Application in Finance, Taiwan Mathematical Society Annual
Meeting.
- Lin, T.Y. et al., Establishing Nanoscale Heterogeneity with Nanoscale Force Measurements, The Journal of Physical Chemistry C, 2015, 119 (32), pp 18267–18277.
- Lin, T. Y., & Liu, M. L. (2013). Index Tracking Models with Forecasting. The 9th International Conference on Optimization: Techniques and Applications (ICOTA 9)
- Lin, T. Y., Chan, P. H. & Liu, M. L. (2013). Men and Women Matching Models and Its Applications. The 9th International Conference on Optimization: Techniques and Applications (ICOTA 9)
