Songyan Hou

zermatt.jpg

Hi, I am Songyan Hou, a PhD student of Beatrice Acciaio and Patrick Cheridito at ETH since September 2021. Before, I completed my master thesis at ETH focusing on machine learning and mathematical finance supervised by Josef Teichmann. I received my bachelor’s degree in Mathematics at Nanjing University.

My current research interests are mainly on Mathematical Finance and Machine Learning. I am interested in exploring models describing the financial world and analyzing these stochastic models. Recently, I am working on applications of optimal transport in mathematical finance and machine learning.

News

Nov 07, 2024 🚀 Excited to introduce Time-Causal VAE ⏱ for robust generating financial time-series!
Aug 06, 2024 🎉 I am thrilled to release NeuralHedge 📈, a PyTorch-based package for deep Hedging, utility maximization, portfolio optimization. NeuralHedge is fully data-driven, lightweight, beginner friendly and flexible. Check it out!

Latest posts

Jan 18, 2025 Gaussian coupling is Gaussian?
Nov 03, 2024 Accidentally commit large files or sensitive data
Aug 03, 2024 Best friends building Python package

Papers

  1. arXiv
    Entropic adapted Wasserstein distance on Gaussians
    Beatrice Acciaio, Songyan Hou, and Gudmund Pammer
    arXiv preprint arXiv:2412.18794, 2024
    PDF
  2. arXiv
    Time-Causal VAE: Robust Financial Time Series Generator
    Beatrice Acciaio, Stephan Eckstein, and Songyan Hou
    arXiv preprint arXiv:2411.02947, 2024
    PDF Code
  3. JMLR
    Instance-dependent generalization bounds via optimal transport
    Songyan Hou, Parnian Kassraie, Anastasis Kratsios, and 2 more authors
    Journal of Machine Learning Research, 2023
    PDF
  4. AAP
    Convergence of Adapted Empirical Measures on \(R^d \)
    Beatrice Acciaio, and Songyan Hou
    Annals of Applied Probability, 2024
    PDF
  5. arXiv
    Convergence of the Adapted Smoothed Empirical Measures
    Songyan Hou
    arXiv preprint arXiv:2401.14883, 2024
    PDF