Skip to content

Lectures and conference materials for the DSE2023 at the University of Lausanne, Switzerland

Notifications You must be signed in to change notification settings

dseconf/DSE2023

Repository files navigation

Econometric Society Summer School and conference in Dynamic Structural Econometrics (DSE2023)

DOI

Topic: Deep Learning for Solving and Estimating Dynamic Models

This public repository contains the training materials, tutorials, lecture notes, and code for the DSE2023 summer school and conference, held at the Department of Economics, University of Lausanne, Switzerland from August 21-26th, 2023.

Anyone is free to use this teaching material. In the latter case, a reference to this repository would, however, be expected

Materials will be added to the repository throughout the course, which will require that students either update their copy of the repository or download/checkout a new copy of the repository.

Citation

Please cite Deep Learning for Solving and Estimating Dynamic Models in your publications if it helps your research or teaching:

@software{simon_scheidegger_2023_8319164,
  author       = {Scheidegger, Simon and
                  Rust, John and 
                  Stachurski, John and
                  Iskhakov, Fedor and
                  Schjerning, Bertel and
                  Yang, Yucheng and 
                  Kubler, Felix and
                  Fern{\'a}ndez-Villaverde, Jes{\'u}s and
                  Miller, Robert A. and
                  Newey, Whitney and
                  Misra, Sanjog and 
                  Sarmiento, Rafael and
                  Azinovic, Marlon and 
                  Chen, Hui and 
                  Malamud, Semyon and
                  White, Matthew N. and
                  Semenova, Vira and
                  Zhou, Kangying and
                  Payne, Jonathan and
                  Chassot, Jonathan and
                  Nuno, Galo and
                  Pelican, Andrin and 
                  Kase, Hanno and 
                  Maliar, Lilia},
  title        = {dseconf/DSE2023: DSE2023 materials},
  month        = sep,
  year         = 2023,
  publisher    = {Zenodo},
  version      = {DSE2023\_DOI},
  doi          = {10.5281/zenodo.8319164},
  url          = {https://doi.org/10.5281/zenodo.8319164}
}

DSE2023 Lecturers, invited speakers, and organizers of DSE2023 (in alphabetical order)

Announcements

We will be using Zulip to post news and links relevant to the event: you should have received an invitation to join the Summer School Zulip workspace.

Online Participation

We will be using Zoom throughout the summer school and conference: you should have received an invitation to join the Zoom space.

Nuvolos Cloud - Our Computational Infrastructure

We will be using Nuvolos Cloud throughout the summer school as computational Infrastructure: you should have received an invitation to join this cloud.

Time table of the Summer School

  • Registration will take plance on Monday, August 21st, from 8:30-9:00 am at Anthropole, room Anthropole 2064.

  • For the Ph.D. student organized hang-out session on Tuesday, August 22nd, from 18:30 onward, please fill out this form

  • For the Pizza and coding event that takes place on Wednesday, August 23rd, please fill out this form no later than Monday, August 21st.

Lecture Date Time Topic Lecturer
01 08.21.2023 09:00 - 10:20 Introduction to deep learning, Tensorflow and, Nuvolos cloud Scheidegger
02 08.21.2023 10:30 - 11:50 Introduction to dynamic structural econometrics Rust
03 08.21.2023 13:00 - 14:20 Numerical Dynamic Programing / best practices in Python Stachurski
04 08.21.2023 14:30 - 15:50 Sturctual estimation of dynamic discrete choice models (MPEC and NFXP) Schjerning
05 08.21.2023 16:00 - 17:20 Stationary and Nonstationary Equilibrium in Durable Goods Markets Iskhakov, Rust, Schjerning
06 08.22.2023 09:00 - 10:20 Deep Equilibrium Nets Scheidegger
07 08.22.2023 10:30 - 11:50 DeepHAM: A global solution method for heterogeneous agent models with aggregate shocks Yang
08 08.22.2023 13:00 - 14:20 Deep Uncertainty Quantification: With an Application to Integrated Assessment Models Kubler
09 08.22.2023 14:30 - 15:50 Exploiting symmetry in high-dimensional DP models Fernandez-Villaverde
10 08.22.2023 16:00 - 17:20 Using deep learning to solve heterogeneous agents dynamic equilibrium problems Fernandez-Villaverde
11 08.23.2023 09:00 - 10:20 Conditional independence and the inversion theorem Miller
12 08.23.2023 10:30 - 11:50 Unobserved Heterogeneity and Finite Dependence Miller
13 08.23.2023 13:00 - 14:20 Machine Learning, CCPs, and Dynamic Discrete Choice Newey
14 08.23.2023 14:30 - 15:50 Deep Learning for Individual Heterogeneity Misra
15 08.23.2023 16:00 - 17:20 High-performance Computing with Python for Deep Learning Sarmiento
16 08.23.2023 18:30 - 20:00 Pizza and Coding Azinovic
17 08.26.2023 09:00 - 10:20 Deep Surrogates for Structural Estimation Chen
18 08.26.2023 10:30 - 11:50 Random Matrix Theory and Machine Learning in Asset Pricing Malamud
19 08.26.2023 13:00 - 14:20 Advances in dynamic programming: theory and algorithms Stachurski
20 08.26.2023 14:30 - 15:50 Endogenous grid point methods and machine learning Iskhakov
21 08.26.2023 16:00 - 16:30 Wrap-up and adjourn Rust
  • Updates of the tentative schedule for the DSE2023 summer school can be found here.

Time table of the Conference (August 24 and 25th, 2023)

  • The tentative schedule for the DSE2023 conference can be found here.

  • Registration will be in front of Anthropole, room 1129.

  • Coffe breaks will be in front of Anthropole, room 1031.

  • Conference dinner will take place on August 24th, 18:30, at the Restaurant du Theatre. Please fill out this form no later than Monday, August 21st, 2023, if you plan on attending.

  • Econ-Ark sponsored reception will take place on August 25th, 18:00, at the Vortex at Le Perchoir Rooftop-Bar at the Vortex building at UNIL Please fill out this form

Prerequisites

  • Basic programming in Python (see this link for a thorough introduction)
  • Basic calculus and probability (Mathematics for Machine learning provides a good overview of skills participants are required to be fluent in).
  • Basic econometrics

Location of the Summer School and Conference

DSE2023 will take place at the University of Lausanne, in the building called Anthropole.

  • You can find all plans of the University bulidings here.

  • This is the Anthropole building:

  • The summer school takes place in Anthropole room 2064 (marked with a red circle).

  • The conference takes place in Anthropole room 1129 (marked with a red circle).

About

Lectures and conference materials for the DSE2023 at the University of Lausanne, Switzerland

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages